Comments on: REBOL to become open source
REBOL Technologies

Comments on: REBOL to become open source

Carl Sassenrath, CTO
REBOL Technologies
25-Sep-2012 8:05 GMT

Article #0511
Main page || Index || Prior Article [0510] || Next Article [0512] || 198 Comments || Send feedback

Update: R3 source has been released. The source repository is: https://github.com/rebol/r3 - as official releases are made, they'll be posted to rebol.com website.

The time has come for REBOL to be released as open source. This is the only way I can see it regaining some degree of momentum and renewed interest -- not just within our REBOL community, but for many yet-to-be users who possess the curiosity or motivation to learn new and powerful programming concepts and techniques.

Here's my proposal

  1. The R3 source code will be released under GPL 2 (most likely - still open to discussion.)
  2. The official source release distribution (rel-src) will be made available from the REBOL.com site.
  3. Developmental sources (dev-src) will be available on GitHub (or a similar service.)
  4. From time to time the dev-src will be reviewed, selected, debugged, polished, and integrated into rel-src releases.
  5. A small group of REBOL Masters (maybe three for now) will be in charge of such rel-src review, selection, revising, debugging, integration etc. as well as develop or encourage the development of new features, ports to new platforms, optimizations, etc.
  6. My role will be to advise and guide such decisions to keep REBOL consistent with its principles, and I will hold the final power of veto, in cases where that becomes necessary.

What I ask in return

If this proposal seems acceptable, there's one thing I ask in return. That those of you who wish to influence and extend the design of REBOL do so with a full understanding of its principles. REBOL is not like other languages, and you will soon discover that REBOL source is not like other source. Architecture and design are important.

Although it is unrealistic to assume we can filter/fix all impurities that might come about in an open environment, we should all strive to minimize the degradation that comes about by seeking an easy solution over the proper and well designed form of such a solution. Even the naming of each symbol, function, and datatype is worthy of thoughtful consideration. That is how I've always treated it.

In the end a language or system should help developers make their programming tasks easier and their products more timely, affordable, manageable, and agile. Too often, it becomes the opposite. Let's always keep that in mind.

What's next?

So... this is it. Once R3 is released our course becomes irreversible. We'll all start paddling our canoes like mad in a small but rapid river... but one which I sincerely hope will carry us to a much larger ocean.

Please offer your comments and advice on any of the above. This is the final call. If you posted a comment on the prior blog that you want me to review again, please note its date and time. I'll go back and take a look.

My schedule is to finish this up next weekend and make the release by October 1st.

198 Comments

Comments:

MaxV
25-Sep-2012 2:27:22
Dear Mr. Sassenrath,

thank you.

You are a lighthouse in the mists of software pollution. When I met Rebol was as an epiphany for me, like the Monnalisa of programming languages.

Now Rebol community will have to face a new Ocean, but you'll remain our Admiral to keep correct our direction.

rebolek
25-Sep-2012 2:46:55
Hi Carl,

that's great news! I have a question, why GPL? You were against GPL in the past (link), so I'm curious why not BSD, BSL or MIT.

Luke
25-Sep-2012 2:47:40
Hi Carl

Thank you for doing this - it will save REBOL from ultimate obscurity. Now I will be able to persuade more people to take it seriously and invest in writing more REBOL code. And I look forward to better integrations with the rest of the ecosystem - such as with XML (its nasty we know, but it's realistically the lingua franca of integrations at the moment)

I agree that any ongoing development philosophy should respect and take forward your architectural choices. Given your massive contributions so far, if you disagree with where it is going, say so, and I'm sure your view will be taken very seriously.

One aspect I think will need to be made clear in the FAQ/documentation is what the licensing status of any executable that links the REBOL library. Will the GPL license conditions apply to tha?. If there would be such restrictions, I respectfully suggest that a MIT/BSD type license might be more palatable and gain better uptake, particularly in the embedded sector. In my view uptake must be a big driver to where REBOL goes now. For example, LUA uses this approach, and is integrated into many contexts.

Anyway, for me GPL is a great start. Good luck with your new ventures, and I look forward to being able to promote REBOL some more.

Thanks

Graham
25-Sep-2012 2:51:14
I'm sure that this has been a difficult choice, but most of us feel that this is the only way forward.
-pekr-
25-Sep-2012 2:53:52
In context of Rebolek's link, I really wonder why GPL (2)? You fear of some bad big company, capitalising on REBOL heritage? GPL is useless, if you push ppl to release the sources to their whole app imo. Not sure, what GPL2 adds to the table. BrianH might have some arguments, how GPL is limiting in context of recent mobile computing work, and that no significant language uses GPL licence ...
Arnold
25-Sep-2012 3:01:31
Thank you for your decision and posting that here.

This means that R2 is stil closed source?! Also it will likely take a lot of effort to enhance R3 to the same level of completeness that R2 has. The knife cuts both ways.

The legal impact of the GPL 2 license chosen at this stage I cannot judge, I'll leave that to those who can.

Good to know you have a firm will to keep REBOL around.

Davide
25-Sep-2012 3:28
GPL2 looks good to me.
I hope that R2 will follow the same fate, because there are many legacy application around to improve and to keep them functional
Mario Cassani
25-Sep-2012 3:30:31
Thank you Carl!
Endo
25-Sep-2012 4:06:04
Thank you very much Carl! Now we can spread the word of REBOL more and more.
Brian Tiffin
25-Sep-2012 5:07:05
Yayys. Great big yayys.

Go REBOL Go

Shadwolf
25-Sep-2012 5:08:40
The time has come for REBOL to be released as open source. This is the only way I can see it regaining some degree of momentum and renewed interest

--> FINALLY !!!

It was that or having a for ever beta rebol2 and a forever alpha r3, not quite the good image you can show to the world.

I hate that your 10 years of investment in rebol ends like "hey dude you know rebol?" "Oh dude, yeah tryed it but it feels so much beta..."

That is great by the way you don't rely anymore on the useless concept of the selected few gurus that have better things to do than drive their tasks to the end.

RobertS
25-Sep-2012 5:11:17
No, not paddling to be sunk in ocean ! (my ancestors often paddled long distances to York Factory on Hudson's Bay.)

Hope to reach a river that has a place to tie up at a BIG dock with a few Showboats in town ! In that canoe? One SOLID soapbox labeled solution: "Rebol3"

Let's see, for some months Rebol3 users looked to be up which creek?

:-)

Shadwolf
25-Sep-2012 5:34:20
The "what you ask in return" Carl is fair. In my opinion of opensourced rebol I never saw it as rebol escaping your supervision and guiding.

You need to focus more on the rebol trend apporting sketch for new technologies to develop in rebol and let the community fill the blank, do the inking and the color :).

You need to be able to think rebol futur rather than doing rebol debugging.

People that are developers knows already that you tightned alot the coding in rebol to make it the most efficient possible even if this implies to sacrifice of funcionality. Porting Gui part is a dilema for example when it comes to adapt inbetwin OS specificities (like transparent background windows... or fonts).

Luca Truffarelli
25-Sep-2012 5:35:11
Great news!

Rebol open source and you acting as a guide for the community.

Thank you

shadwolf
25-Sep-2012 5:41:38
R2 have to be opensource too... there is to much good things in it that you let down to focus on r3 like rebcode, rebservice, rebwebplugin etc etc ... At least to be the set mark of rebol 3 and as rebol 3 is modular adding rebcode as side module could please more than a ton of people.

Tons of great great ideas that has never got a chance to get through due to lack of time and overwhelming project.

shadwolf
25-Sep-2012 5:45:27
as for modules to r3 I liked the rebol/browser and it's capacity to edit and difuse in an easy way the collection of rebol scripts. Why not thinking something alike for the rebol extensions (the techs exists look at the ubuntu software center...)

I think the goal of the rebol community will be to do the maximum of things in rebol and hack their ways in C only when they face an limit that forbid them to continue in their all rebol implementation.

Steven White
25-Sep-2012 6:03:52
Thank you. That is good news for me.

Where I work, we are looking for some new programming tool to use in place of one we used to use but that is fading in popularity. The argument against REBOL was that we don't want to write a lot of code in a language that is doomed to die off. Now I can argue back that it probably is not going to die off and its relative simplicity makes it a good choice in contrast to some alternatives.

I hope that R2 will stay around for a while, at least until R3 matches its capabilities including VID. I don't write a lot of REBOL, but I do have a few real production applications in R2 that are in daily use.

Interestingly, but off topic, at one point I had managed to get the free (as in beer) REBOL interpreter into our standard disk image, meaning that whenever we deployed a new computer, it had REBOL on it, ready to go for whatever use I could think up for it. So there are people out there really using your creation.

Giuseppe Chillemi
25-Sep-2012 6:05:04
Great Carl ! Please, prepare a clear plan about the furure of REBOL (Datatypes, devices and so on) and the community will implement these features.

Thanks, Giuseppe Chillemi

Nick
25-Sep-2012 6:09:23
Thank you Carl! I think it was a necessary decision. This announcement needs to be posted in many developers' forums - let the evangelizing begin :)

I will echo some of the same sentiment so far: BSD or MIT style license may get REBOL more traction, and there would certainly be benefit to having the code in R2 released so that the huge volume of legacy REBOL code can be moved to future platforms.

shadwolf
25-Sep-2012 6:37:31
CArl and by the way I tooooooootally understand your point of view "too many cooks ruins the dish" this is why the developement of opensourced rebol can't be set without supervision but no projects opensource works on total out of any control mode.

examples are many Gnu CC ? python? perl? tons of cooks big dish.

there is a field where a big number of people is absolutly requiered and it's the debug process. Is it better to have a bug database filled to the skytop and one single event extremly talented to write the patch or is it better to have 10 000 user proposing source code fix for each bugs they run through..

Brian Wisti
25-Sep-2012 7:30:57
Hey, GPL may not have been my first choice, but only because of my current job. I'm still ecstatic that REBOL will be released open source.
Eric H.
25-Sep-2012 7:31:13
Thank you Carl! I've been wanting this for a while now. I see great things happening with REBOL now that the source code will be available. I'm no longer a "REBOL Master" but with this my usage will increase again. I'll contribute what I can to the project though at this point some of my ideas may be obsolete. I'll have to go over some of my notes.

Thank you again.

Eric

Fork
25-Sep-2012 7:35:34
Great news!!

I was/am in the minority (of readers here, emphasis mine) in thinking copyleft will do more good than harm. But it's a minority of at least 2 now. :) This clearly loosens the grip of control, and I'd have thought the community would only want to share with those who plan on at least offering their changes to the interpreter back! (You don't have to take them.)

Arguing a GPLv2 system can't gain traction in the embedded market is a straw man, given examples like...Linux? Mentioning version 2 *explicitly* suggests the license will almost certainly contain something like the tweak Linus threw in:

"Also note that the only valid version of the GPL as far as the kernel is concerned is _this_ particular version of the license (ie v2, not v2.2 or v3.x or whatever), unless explicitly otherwise stated."

This draws Stallman's ire, and there is probably a place here or there that matters beyond ideology. (Only one I know of is exclusion from GNU's Savannah source archives, which is a dinosaur compared to GitHub anyway.) Yet it should give peace of mind to companies that accept the "publish your modifications" part, but don't want their product's behavior dictated by revisionist license changes that come after that [such as Tivo's modifications].

Hopefully we won't have to wait too long for the source! (Please?) If any help with Git is needed, I'm happy to offer. I've learned a bit lately, including even alternative workflows such as Gerrit (which is quite different from GitHub, and I wouldn't suggest except for much larger scale projects like Android). But I'm sure some here know more than me about Git.

Beyond versioning...there's been longstanding Rebol-based code for communication, documentation, and issue tracking. Hopefully these codebases count in the "open-source" wave. Still, it might be worth evaluating other tools that can be brought in to help. That's even if they are orders of magnitude larger in megabytes--but only if they're actually more useful!

I'm looking forward to seeing what can happen. Once the source is out, it might bring people back and turn the evangelization engine back on faster than one might expect. Just saw a note on G+ from Douglas Crockford from last week about the previous blog post:

"It has been a long time since I have written in REBOL, but I continue to admire the language and appreciate the things it taught me. JSON would not have happened if REBOL had not prepared me to recognize it."

-pekr-
25-Sep-2012 7:55:40
Fork, I thought you would be happy :-) I know that Red is kidn of a competing product, althought a bit complementary one. But here's what DocKimbel thinks:

"The GPL precludes me from looking at the code, the risk is too high to unconsciously write similar code and infringe the license. It is even worse than that, all Red contributions implementing a feature that exists in R3, will need to pass through a peer-reviewing process to determine if it is a derivative work of R3 or not. The reviewing persons would be able to look at R3 sources, but doing so, would not be able to contribute code to Red. So I strongly advise current and future Red contributors that wants to add REBOL features to Red from looking at R3 sources.... Porting GPL code to another language falls under derivate work clause too. If it wasn't the case, it would be too easy to workaround GPL terms by porting it from language A to B, then from B to A .... There is a reason why most big software companies prohibit their developers from looking at GPL source code."

So - will it lead into a community split?

Kyle Patrick
25-Sep-2012 8:02:55
In my view, languages need to be as open as possible and that means MIT. The structure of REBOL (like Smalltalk and FORTH) means it is highly dependent on it libraries, which argues that the libraries need to be MIT if the language is to succeed.
Greg
25-Sep-2012 8:03:21
This is clearly a second chance for REBOL to spread as much as possible. A new start. Good move.
Fork
25-Sep-2012 8:34:49
Well...if DocKimbel sees a neat snippet of rebol.c he'd like to borrow for Red...what if he asks if he can have that snippet and license it more liberally? By choosing a conservative license, Carl gets to make the call on a case-by-case basis. It can depend on who's asking, and what they're asking for. :)

As long as you're not copy-pasting you're fine either way. The philosophy of GPL code is not "God forbid someone *look* at it and be inspired to make a more liberally licensed codebase in the same spirit". Ideas simply don't work that way, and Red was doing *that* before this announcement. GPL is only designed to stop someone using another's shared work out of whole cloth--passing the product on to someone else--and not giving the person you passed on to the same rights that you had.

...and the copyright owner can make other arrangements, paid or gratis, for whatever additional license they want. How much more fair can you get?!?

Shadwolf
25-Sep-2012 8:49:35
The good question isn't around the opensource licence type.

It's more about will we use that opportunity and does the world will notice finally rebol?

Serriously I prefer 10 forks of rebol advancing on their own pace proposing their own set of goodies and living and dieing like biological organisms (that you can see with the linux distribution frenzy how many linux distro where created and shared since the linux beginings? How many of them stayed in the history and truely impacted their sourroundings?) than one rebol still in alfa and merely a concept.

Pekr the community is already split no ?

What can really happend to red and that was an early projection I made is that the ex-reboler that went to red returns to rebol now that they can have an impact on the real thing.

if red is really mandatory it will goes on supported by not only the ex-rebolers. If red was just a hobby it will disapear like the other rebol inspired projects. most of the things that community ever produced where fated to disapeared since based on a lone hobistic effort.

Fork
25-Sep-2012 9:01:17
Red brought a new "systems-level" dialect. No such thing existed in Rebol, and experiments with dialects were always needed to explore their potential, even Rebmu :)

There's no reason to abandon the effort. Maybe the change is just that Red doesn't feel so eager to drop Rebol from the bootstrap process--and can accept it as a part of the permanent toolchain.

Kaj
25-Sep-2012 9:15:54
Stay tuned.
Fork
25-Sep-2012 9:17:28
For perspective, even Yukihiro Matsumoto (a.k.a. "Matz", the chief creator of Ruby) cares enough to tweet about this announcement:

https://twitter.com/matz_translated/status/250614360794333184

So don't lose this momentum, publish sources soon! I wouldn't worry too much about making it foolproof in terms of build, first responders will take care of that and patch quickly. People will forgive funky build problems, just not a long delay from the announcement. :-/

Brian Hawley
25-Sep-2012 9:23:27
Unless the license has at least a Classpath-like exception that reflects the way that REBOL code is integrated (hint: not by linking), I won't be able to look at the source, let alone make any improvements. I won't even be able to use the source function. R3's source is now more closed to me than it was before.

My existing contributions are still MIT licensed, including R2/Forward.

Nathan
25-Sep-2012 9:28:50
I hope you consider using a more permissive license such as MIT. Using GPL in a a language or a framework simply ensures that a very vast swathe of programmers will be unable to use REBOL and will ignore it. For example, see the discussion about meteor.js at Hacker News:

http://news.ycombinator.com/item?id=3836212

Eventually meteor.js relented and switched to the MIT license; I hope you'll do the same!

yo
25-Sep-2012 9:34:50
Me parece una gran decisión, aunque se debería evaluar mejores licencias como la MIT o BSD.

Los conceptos y arquitectura me parecen grandes, pero hace falta mayor información y divulgación. Espero que con esta apertura se tenga mayor colaboración y uso.

Carl Sassenrath espero que sigas aumentando las ideas computacionales y el mundo libre las magnifique.

Andreas
25-Sep-2012 9:35:32
Carl, thanks a lot for coming to a decision and for communicating it clearly. Overall, the plan sounds good and sensible. One but:

Please consider to use MIT/BSD-like licenses, or at least the LGPL instead of the GPL.

Under a strict interpretation as championed by the FSF [1], R3 under the GPL will require all scripts written for this R3 to be GPL-licensed as well. Hardly any R3 script would be able to work without using GPL'ed mezzanines (such as FUNC) -- per the FSF, a program/script needs to be GPL'ed if it calls interpreted libraries that come with the GPL'ed interpreter.

[1] https://gnu.org/licenses/old-licenses/gpl-2.0-faq#IfInterpreterIsGPL

Jos'h Fuller
25-Sep-2012 10:01:25
Aaaaahh!

Thank you!

We could have no expectation that you would do this, and there was no obligation for you to do so. I hope the community will be able to repay your hard thinking and tough decision with blooming flowers of awesome!

I have been setting up an application on the Rasberry Pi. At every turn I was thinking "oh, but VID does this so much better..." The idea of having REBOL working is completely inspiring!

(N.B. I really hope REBOL could strike quick to become the "BASIC" of the Rasberry Pi. REBOL is so much easier to deal with than the mishmash of stuff required for other scripting languages. I am certainly ready to help with compiling, testing and debugging!)

Did I mention? Thanks again, Carl.

Brian Hawley
25-Sep-2012 10:25:33
Andreas, it's worse than that. The only parts of R3 that are actually interpreters are do and parse. Every other function in REBOL is a library function, including the natives.

On the plus side, as long as you don't do something at all similar to encapping, or make your own host code, or make any extensions, or use any closed-source extensions, you can still run scripts since scripts currently either come in source code or compressed source code. We never got around to adding support for encrypted scripts, so you're basically distributing your source anyways.

Host code for a plugin that is meant to be loaded by a closed-source application, or extensions that wrap closed-source libraries that weren't shipped with the operating system, you won't legally be able to make those. You can't even integrate with or run on .NET or Java because that wasn't legal until GPL 3 (yes, that includes all existing GPL 2 Java applications), though you can run on Mono or OpenJDK.

No iOS or OSX App Store applications. No using Google-specific or other proprietary libraries on Android. No WP7 applications. Not sure about Windows Store.

Bo
25-Sep-2012 10:42:29
Jos'h Fuller, I am also working with the Raspberry Pi, but also despise the mishmash development environment.

I think Rebol could easily be the de facto language for RPi.

BTW, I just sent a msg to Nixie Pixel of OSAlt to request her to do a piece on Rebol going Open Source. If you want to help get the word out, maybe you want to stop by her Contact page and let her know about it. http://www.nixiepixel.com/contact/

Sending her a little love using the "Send Nixie Love" link on her page might help sway her as well. ;-)

Along these lines, if you know of any other news or social media outlet to spread the word, post it here so we can all send requests for coverage.

Brian Hawley
25-Sep-2012 10:49:14
Fork, in order for Carl to grant permission to Doc to use some part with a more permissive license, R3 would need a FSF-style copyright assignment policy. Otherwise, Doc would need to get permission from the copyright holders, if they can be tracked down, similar to the situation with the Linux kernel.

Also, if Doc "sees a neat snippet of rebol.c", he is legally contaminated already. Someone else who is not contributing to Red or Topaz would need to see the code, then get the permissive license, and only then would Doc be able to see it.

You can replicate behavior when cloning, but you can't look at the source. If the behavior is patented you can't even do that.

Joe Latone
25-Sep-2012 10:52:39
Hi Carl,

We met many years ago, I came across this news recently, and a few comments:

1. I'm most excited just to see your code! I always like to see the source code of a craftsman.

2. The timing couldn't be better! With Facebook's miss on mobile, the time is ripe for this, http://thenextweb.com/facebook/2012/09/25/66-mobile-developers-think-facebook-will-disrupted-social-app, and no better way to build a p2p mobile-first, secure social network than with REBOL, yes? You've been doing it for years.

3. The GPL's going to discourage commercial endeavors, especially if it covers user-written REBOL scripts. You could instead use the EPL, a "weaker" copyleft, which would still require contributing back to open source any changes or derivative works of REBOL itself, while still allowing qualifying extensions to the source code as well as REBOL scripts to be licensed in any way the author chooses. The Eclipse community is a thriving commercial AND technical success, and REBOL could be, too. It's still that far advanced! Otoh, if you want to discourage commercial use and/or offer a (dual) commercial license, then the GPL is the way to go...philosophy discussions about "freedom" aside.

4. You might wan to consider taking copyright assignment on any contributions if you go GPL. Why? It's going to be hard to change your mind on the license later--either another OSS license or a commercial license--if you then don't own 100% copyright to the code.

Best, Joe

P.S. I'm not posting as one, but I am an IBMer. I'm disclosing this in case people think I have a bias towards "commercial" and the EPL because of IBM. I don't. These are my personal opinions. I also like the GPL, MIT, Apache licenses, but I chose to argue for the EPL because it matches what I'd like to see going forward with REBOL. :)

Brian Hawley
25-Sep-2012 11:36:42
EPL doesn't look that bad. It would exclude cooperation with Boron and Orca (they're LGPL), but not necessarily with Red and Topaz (they're MIT). Classpath code should be OK as long as it's in separate files from the EPL code. If there's a similar license that's GPL-compatible we might want to use that instead, but only if it's not worse. Anyone experienced with MPL 2?
Steven White
25-Sep-2012 11:48:58
This licensing stuff gives me a headache, and I'm sure I don't completely understand.

It does indeed seem that if one writes a script to be interpreted by a GPL interpreter (REBOL), and that interpreter provided "bindings" to "facilities" (the mezzanine?) that are GPL, then the script is technically "linked" to those "facilities" and therefore must itself be released under GPL.

However, is it possible that the GPL was written with a different paradigm in mind, in the days of compilers and linkers that operate on "object code" produced by compilers? Is it possible that REBOL, under the hood, is different in some ways that the GPL authors didn't think of? Would it help if Carl actually called the Free Software Foundation and asked them (617-542-5942)?

I don't pretend to understand this in detail, so this might sound like a laughable idea, but it seems to me that the spirit of the law, if not the letter, should be that it is possible to write a closed program using a GPL tool, the way one might write a closed C program and use the GNU compiler (if that is possible, I guess I don't actually know).

Steven Solie
25-Sep-2012 11:51:48
I think it is the only move you can make but I am also cautiously optimistic.

REBOL is great because it is well thought out. I just hope the "Jedi High Council", or whatever it will be called, will be able to keep control of R3 yet still gain enough momentum from contributors to flourish.

In any case, I am very much looking forward to finishing my work on the AmigaOS host kit. I was blocked due to some bugs in the core which can now be addressed.

This brings us to the question of the license. I think you need to put some more thought into how exactly the licensing is going to work before going public. I believe any misstep could result in a loss of developer and commercial interest. For example, will all the host kits be forced to be GPL if using static linking?

I am also curious about any REBOL trademarks. You will need to iron out some kind of policy there.

Finally, I wonder if you have any software patents which may need addressing before going open source. The last thing we need is for R3 to go open with some patent troll quietly lurking in the background.

Joe Latone
25-Sep-2012 12:20:33
On the subject of the GPL and REBOL scripts: Carl can put the subject to rest by merely declaring that the REBOL engine itself is GPL but REBOL scripts can be licensed in any way the author chooses. This is no different than Linus & the Linux kernel (see the statement from Linus about the kernel & userland in http://www.kernel.org/pub/linux/kernel/COPYING) or Larry Wall & Perl (see the statement from Larry about the Perl interpreter and Perl scripts in http://dev.perl.org/licenses).
Brian Hawley
25-Sep-2012 12:26:03
Steven W., you can make closed-source programs with a GPL compiler like GCC, as long as you don't include any GPL libraries. For instance, Cygwin GCC normally links the cygwin library, which is LGPL, so the programs it generates have some restrictions. The Mingw GCC compiler doesn't normally link to a copyleft library, so the programs it generates are less restricted.

Steven S., if R3 is GPL, apps made with host kits would be required to be GPL even if they are linked dynamically. LGPL allows dynamic linking, or relinkable static linking. Classpath allows static linking without requiring it to be relinkable.

Shadwolf
25-Sep-2012 12:45:40
hum Fork I would lie if I say that opensource rebol isn't something every people that looked somehow into rebol ever wanted.

It's great that some hours after the announcement the information spreads at high pace and people feel their interest for rebol reborn.

The contrary would have been so sad ...

Shawolf
25-Sep-2012 12:51:32
as for Red's futur. I think that if it's forbiden for the Red devteam to participate in the enhancement of r3 opensource to not infrige licence regulating we will split the effort and most of the most talented devs will not participate in the futur adventure long await of rebol3 as opensource.

Is that good? Does those people that participated and kept faith in rebol even in the darkest hours and tryed to set things in other ways can pass by that oportunity with smug face?

Seeing that I'm happy that I didn't joined the red devteam cause my priority was and remains rebol.

Why not seeing red as the next r3/sdk ? Why not give a mutual futur to both project puting the best efforts and gathering the forces to make both things big ? Each of them as a purpose therefore each of them should exists.

Joe Latone
25-Sep-2012 13:04:40
My final word on OSS licenses: Carl owns the copyright. He can set the terms of the license, what's covered by it, and what's not. For example, there's GPL software out there where the author specifically declares that the output of the GPL software is covered by the terms of the GPL as well, in case anyone had the notion of coupling the GPL software with non-GPL software via IPC, file IO, etc, to keep the GPL at arm's length to try to avoid being encumbered by it.

What's important is that Carl be explicit about it all, so the community knows what it can do, esp with the GPL, so nobody has to play or pay an IP law attorney and guess about something that hasn't been tested well--if at all--in court. And this is especially true if REBOL's going to be accepting contributions.

It's the author that matters, not only the license.

Peter F
25-Sep-2012 13:21:41

Using the GPL 3 would protect against Tivoization which the GPL 2 does not.

If REBOL is suitable for embedding into large code bases, the LGPL would allow this, while the GPL does not.

LGPL 3 would provide the best of both worlds.

Pierre
25-Sep-2012 13:24:02
T H A N K     Y O U  !

M E R C I !

Sure, licence matters. A good licence needs to be carefully chosen: many advices were given, Carl and RT must make up their mind, peacefully.

But the main message is just wonderful: freedom.

I'm going to dig out my old rebol code. Now I know that it cannot die.

Vive la Rébolution!

Steven Solie
25-Sep-2012 13:57:48
Brian Hawley, my GPL query was just one small example of the questions that will be generated if the whole licensing thing is not well thought out. I once checked with an FSF lawyer and he said the opposite (i.e. dynamic linking is OK but frowned upon). This stuff can easily get quite messy.

This is why I feel it is wise to consider all the scenarios and their possible commercial implications before releasing the sources. Changing licensing terms after the fact is possible of course but I'm not sure that is a good approach.

bdug
25-Sep-2012 13:58:24
"A small group of REBOL Masters (maybe three for now) will be in charge ..."

This sentence worries me. Does it mean there will be only one rebol project? Will free forks be forbidden ?

Several rebol forks would lead to competition between them and foster innovation.

I fear that only one rebol project would make rebol to progress too slowly.

Graham
25-Sep-2012 14:06:24
I think it means that there will be one official branch sanctioned by Carl. Forks will be unofficial.
Jonathan Abbey
25-Sep-2012 14:16:26
Wonderful news, Carl. I've known that REBOL existed for many years, but there seemed no big reason to mess with it when I could use so many open source languages and be sure that anyone would be able to use my code.

If nothing else, REBOL looks like the language I'd choose to teach my daughter programming with, and if it becomes open, perhaps folks will do the work to port it to Android, to provide 3d support, media handling, and etc.

Congratulations and thanks!

Gregg Irwin
25-Sep-2012 14:17:20
This is exciting news. While we can argue licenses back and forth, a small doc by someone like BrianH, Kaj, or Andreas (anyone who *really* understands the issues and implications), seems best. If it spells out what you can and can't do, or what you have to release under different scenarios, we will each know how we can use it or if we will be contaminated by looking at the source.

I would prefer BSD/MIT so there is no confusion, since I use REBOL in commercial work. My larger concern is how to support and leverage common mezz level code and libraries across REBOL-like languages without licensing confusion.

Gregg Irwin
25-Sep-2012 14:23:40
It could also be instructive for Carl to note key reasons for choosing the GPL2 license.

In any case, it's a big day for REBOL. Thank you Carl!

Nick
25-Sep-2012 14:34:43
I love that Matz already tweeted about this. Tiobe, here we come :)
Carl Read
25-Sep-2012 14:56:11
Carl, regarding the license: To be successful, post-opening, REBOL will need a good group of core programmers working on it who, as you say, understand its principles. To attract and keep those programmers will require a license that's attractive to them. So I suggest you poll those who were working or R3 (who in your eyes get REBOL) as to which license they'd most prefer. They're the ones who'll make or break REBOL now, so you'll need them on board.
Fork
25-Sep-2012 15:23:17
(a) Carl's stated goal is to open Rebol up.

(b) The previous license terms allowed you to transmit official unmodified binaries with your source.

(c) People have successfully released Rebol projects and libraries under the license of their choice, for over a decade.

How—exactly—would it be becoming "more open" if people could not do at least (b) and (c)? Perhaps I'm jumping to conclusions here, but I'd say the intent is almost certainly:

"IF you change any files in the official distribution AND you transfer this to another party, you must also give that other party the source changes."

Maybe a tweak or two is needed beyond the GPLv2 or LGPLv2 to get this intention in clear "legalese". But I don't think the goal is to restrict anything about what kind of programs you can write on top of it if you leave the runtime + base libraries unmodified (or if you share your modifications to the runtime + base libraries).

There are a few administrative issues to watch out for. If the main branch wants to keep its re-licensing options open, it has to go through a contributor license agreement. This happens with contributions to other projects like Qt...and if you don't agree to those terms (or don't have the right to surrender those terms) they don't take your changes. This doesn't stop you from distributing your mutations if you've published the source, but it does decrease the likelihood they will become canon.

Ben Scherrey
25-Sep-2012 16:31:29
Respectfully, GPL is a non-starter for me. Actually likely to result in the opposite of your intentions as the only real choice many people will have will be a re-implementation and that will never be done with the same sense of priorities and values as the original. Language will split, flounder, and die.

You've got the perfect example of how to build a community and take a language forward with Python. I hope you take it and REBOL gets the chance to move forward that it deserves.

Best of luck!

Alex A. Naanou
25-Sep-2012 17:04:20
IMHO a tool (language, libs, etc.) should be at least LGPL, this will enable maximum freedom to the user while still providing some strings to pull for the owner.

I'd go for BSD/MIT if it was up to me, but that is just my opinion, LGPL will do allot better than GPL ;)

Thanks!!

Chris Travers
25-Sep-2012 18:24:40
Anything but the GPL v3 ;-) I am happy with either the GPL v2 or MIT/BSD.

This is absolutely great news. I would like to say one thing though. For those who are worried about being in an open environment, you may want to look at how other well-engineered projects run things. Examples I would look at are some of the *bsd communities and PostgreSQL. Beign in an open environment doesn't (and shouldn't!) mean less control over the architecture.

I will probably try to follow up over email and offer some (advisory) help with the transition.

Brett
25-Sep-2012 18:42:48
Carl,

Going open sounds great.

But perhaps you should get a commitment, if you have not already done so, from the people that will form the small group of REBOL masters that they are eager to work with your proposal especially the license part.

I'm not up with the intricacies of the GPL because I usually avoid it, but if Brian's comments are correct then I'd be pretty dissappointed with the implications he's already outlined.

Jerry Tsai
25-Sep-2012 18:55:05
+1 to Carl Reed
Jaime Vargas
25-Sep-2012 19:12:34
I recommend using the Open Source Initiative OSI - Common Development and Distribution License (CDDL-1.0) http://opensource.org/licenses/CDDL-1.0

Files licensed under the CDDL can be combined with files licensed under other licenses, whether open source or proprietary.

WuJian
25-Sep-2012 19:22:14
the more freedom, the more support from good programmers.

the more killer apps from REBOL, the more success of REBOL.

the more REBOL brings to the world, the more REBOL brings to the BDFL.

W^L+
25-Sep-2012 20:24:09
I am a fan of the GPL, but I believe that the LGPLv2 or Mozilla Public License will better meet the objectives.

For example, I believe that it should be clear that the chosen license does not affect the license of any library or script used or produced with it. On the other hand, Rebol-alikes that utilize official Rebol code should be produced in such a way that changes may be pulled into the official Rebol distribution.

The license should allow (and encourage) more people to lend a hand, either in the official Rebol distribution or in a forked Rebol-alike distribution or in libraries / add-ons / extensions for Rebol. Hopefully will happen without the parasitism that BSD tends to encourage.

The license should encourage ports and experimental forks. This is the only way that we'll see Rebol on Android and iOS, for example.

The license should encourage widespread usage of Rebol "in the wild"

Finally, the license should make it clear that commercial and fee-based use / libraries /extensions are allowed and encouraged along with those which have zero-price. Again, the license of Rebol should be specified as not affecting or being affected by the license of any such use / library /extension.

I think that Mozilla, in formulating its licensing, had to address these same issues. I think it will be hard to beat MPL (Mozilla Public License).

Brian Hawley
25-Sep-2012 20:40:59
If we keep the same system model, we can do quite well by licensing r3lib.dll and r3.exe as LGPL, but the host code as MIT or something else more permissive than the LGPL, maybe Apache.

You would then be able to make extensions that can wrap closed source DLLs that weren't included in the operating system you're running on, but we would still be required to use r3lib in a relinkable form in our host code apps. You could even make hosts that are plugins for closed-source apps, like Visual Studio.

We could also use a FAQ entry or something that explicitly says that encapping, loading and binding REBOL code is not linking, it's just loading data. It's just something to reassure your users' lawyers :)

And you could still offer commercial licenses that would allow you to link r3lib statically in a non-relinkable form. That would require copyright assignment, but that's expected.

Brian Hawley
25-Sep-2012 20:49:58
It would be helpful if the mezzanines were permissive too, or else use of the source function would be restricted. You've licensed a lot of the mezzanine code under the MIT license anyways (from me, since I didn't do copyright assignment), so it wouldn't hurt much to license the rest with the same license.
Bo
25-Sep-2012 21:28:25
Carl, PLEASE also open source R2 so I can at least try to build it on the Raspberry Pi running Debian on an ARM v6 processor. I have at least one order for a commercial product I'm developing in R2, and the Raspberry Pi is a highly-valued component of that system.

Or, can I come over and try to build R2 on my Raspberry Pi in your basement? ;-)

James_Nak
25-Sep-2012 21:42:48
Carl,

Arigatoo.

Maxim Olivier-Adlhoch
25-Sep-2012 21:54:46
I'd really like to play around with R3, but if its going to be GPL, I'll be-grudgingly be forced to ignore it.

I can't add my own software into it and be forced to license it as GPL. LGPL would probably allow me to USE R3, but not contribute much to it. That's because any stuff I add to R3 I can't add back in my own software.

furthermore, if I unintentionally copy/adapt a tiny bit of code I see within the source within another project, then I'll be forced to relinquish it, even if its not an actual copy of the project in spirit, but some bogus and arbitrary code bit.

GPL is a noble idea. But unfortunately, In the real world, we live by selling products to clients. Open sourcing languages, libraries and APIs is nice, but not when it forces me to give away the very work which feeds my family.

All code I publicly release uses MIT licenses. A large part of it was payed for by commercial customers and they agree to use MIT licensed tools, because I'm not forcing them to also release their proprietary software. If I where, they would not use my code and even less allow me to release all of it.

The idea that we should all give away our livelyhood may be possible for academia types, and software may be secondary when you sell hardware, but not when its the product you are actually inventing and selling.

Until I can go into a supermarket, a car dealer and a realtor, and get their products for free, I can't play around with GPL tools which force me to give my products or my employer's for free.

Arnold
25-Sep-2012 23:17:54
Maxim has a point here.

This is the reason I have chosen to be "on the Red team", (although I have not yet produced a single line of code) when the R3 license stays GPL(123) (and if Nenad says that is ok). Because Red is REBOL strongly inspired I guess it is allowed to look at the sources of REBOL functions using source function. This is inevitable anyway because this is how we learned REBOL and by viewing many of the scripts Carl and the RT-team have provided over the years.

I can see Red having many benefits over REBOL even in a near future. I call on for a codesprint for Red and ask Carl hereby to please postpone open sourcing R3 to the first of januari 2013.

Fork
25-Sep-2012 23:52:10
I hear your point of view, Maxim. But my lifetime--whatever it is worth (perhaps not much, to this crowd)--has hinged on seeking to unlock the resource constraints that I grew up with.

No - right now we can't simply go into a store, or car dealer, or realtor and get a free House/Car/PizzaBagel. But I've learned enough about brilliance and blockage to know that the limit is only if "we" (I'll use the term loosely)--the elite software vanguard--can't stand by the foundations of unlimited potential that we know to exist, rather allow it to fall into the wrong hands.

When I was a kid, I thought records and tapes were a limited resource. Time and experience taught me about the nature of sound, the encoding and recycling of bits and generic representation. I feel sorry for all the great musicians whose voices were never heard because of the high cost of the instruments, or who were sold department store toys and blamed themselves because they'd been marketed "amazing" products.

(I'm looking at you, Sears "100-voice" Yamaha keyboard...)

The digital age brings us within reach of freedom. And it's robber barons like Steve Jobs and the DRM movement who've wanted to slam the wall down again so everyone must start again. Perhaps we're even worse off than before, because the lockout complexity is increasingly artificial and legal-patent-based (vs. "nature" based..or were the "natural" limits a result of a previous iteration of shortsightedness?)

Perhaps you've got your eye on a certain Porsche or property on the beach--and if so, I'm quite sorry to speak cross-purposes if none of this means anything to you. I just don't want children to walk penniless into the world and be given Microsoft Paint, while gullible developers long dead gave "institutions" like Pixar their toil for false dollars and control of the societal narrative.

Additionally: GPL and MIT are compatible if you would try to bridge your customers' concerns:

http://stackoverflow.com/questions/3336161/mit-gpl-dual-license-in-commercial-software

I'd like more of the "random person" naysayers here to speak to the specifics. My (a) (b) (c) simply must be on track with Carl's agenda. So paint a clear scenario for what you want to do that can't be done under the *assumed* license. If you can't express your closed-source idea as a dollar value for the contract you want ($1000? $10,000? $1?), then why should Carl or anyone be panicked over your wacky valueless customer scenario?

And as for iOS, Apple is rotten to the core. Look...they kicked Google Maps out merely for being a competing product to their own - no interpreter, no GPL, just "we have a map product so we're dropping yours". Holding your license choice hostage for an organization that pulls that AND charges you $100/year for the privilege ... well, people stay in abusive relationships for some reason. (Also, catching up to Objective-C for iPhone is going to be hard, even if anyone actually cared.)

Fork
26-Sep-2012 0:08:07
(at Arnold) I updated my answer to the closed question "Is Rebol Dead?" on StackOverflow. (The technical nature of the site is that even if no new answers can be added you can edit existing ones, it's just how it works:)

http://stackoverflow.com/questions/7577186/is-rebol-still-alive/7578438#7578438

Though I'm less interested in the concerns of a compiled Rebol variant, I'm still curious about Red. In my update I gave Red credit for its progress, and I acknowledged that it's a bit unfair for this decision to be announced right when it is coming into its own.

But there's no reason this announcement can't be good for both projects. At worst it's a no-op, but it doesn't have to be. Neither is benefitted by delay in any case.

Robert M. Münch
26-Sep-2012 0:53:31
Please use LGPL or even better MIT / BSD but not GPL, since with interpreters it becomes to complicated when doing commercial things.

With GPL Rebol is locking out the users that can afford and sponsor further development.

If you want Rebol to spread, you need people willing to push it forward (invest time and money) otherwise it won't happen.

Giuseppe Chillemi
26-Sep-2012 3:11:44
I have read everyone concer about licence. Go for MIT/BSD !

Maarten Koopmans
26-Sep-2012 3:15:43
....agreed on the licensing objections. I'd rather have closed source than GPL. Robbert makes (as always) a sensible suggestion with LGPL.

LGPL will protect you - whilst allowing people to use it commercially.

TomBon
26-Sep-2012 3:58:26
+1 Maxim/Robert
florin
26-Sep-2012 4:18:17
Now we're talking. Forgiven.
Shadwolf
26-Sep-2012 4:45:56
For the licence issue instead of poluating this thread with rant about it please express your self in this poll.

http://gopollgo.com/what-licence-do-you-want-for-rebol-opensource

Vote massively this is now that your opinion count later it will be too late.

shadwolf
26-Sep-2012 4:58:15
I swear wujlan isn't me using another nickname... He nailed perfectly the thing. it's like 10 years I say: No big ambitious apps made upon rebol (like a PHPBB or an OScommerce etc...) no way to attract people. Big ambitious apps that fill a vital need is the best way to show the rebol ways and give a large scale point of comparaison betwin rebol and the rest of the world. For the moment the apps we have made in rebol are great but are still far from a visual studio for example. You can say all you want about MS visual studio but it is the most advanced IDE and it looks ultra pro. Having such a high quality IDE for rebol and red project management testing and deployement would be something amazing (yes I preach for my church, so what :P?). because an IDE is the most accurate recollection of all the rebol specific technologies. Parse and Draw/agg for the text coloring, net stack for deployement publishing and documentation, GUI used in extensive way with widgets like tables, treeviews, menus, drop boxes, dockable items, tool bars etc...
Alexander Boström
26-Sep-2012 5:04:40
New to Rebol, but some suggestions:

1: Really that GPLv2+ ("GPLv2 or later"). It would be a shame to not be able to combine this work with GPLv3 code. Once you decide on GPLv2 only it will be very hard to change that.

2: Consider adding the Classpath exception, as used by the OpenJDK Java implementation. This is a way to allow people to build on top of Rebol and making the copyleft only apply to changes to Rebol itself. This is probably more useful than choosing LGPL.

https://en.wikipedia.org/wiki/GPL_linking_exception#The_classpath_exception

Jaime Vargas
26-Sep-2012 6:05:03
If not MIT/BSD, CDDL because in contrast to any GPL based license, it does not have the virality problem, so there is no problem with reusing the code in proprietary software. At the same time, it has all the advantages of copyleft license (it's just a slightly tweaked Mozilla license). It is easy to read and understand, under two pages. Not only covers copyrights but also patent claims. It was used to open source a large highly encumbered proprietary code base, the Solaris Operating System.
Alexander Boström
26-Sep-2012 7:23:59
If the goal is to renew interest in the software I think it's helpful to select a GPL compatible license, which the CDDL unfortunately is not.

GPL with Classpath exception is GPL compatible. LPGLv2+ also works. Apache License 2.0 is another popular option. The Perl dual-licensing scheme works. Mozilla code (Firefox etc) is triple licensed as MPL/LGPL/GPL which makes it GPL compatible. And then there's BSD/MIT if you want to keep it simple.

Dmitry Northerner
26-Sep-2012 8:43:18
It's great news for all of us!!! But, license, why not BSD or MIT...
Steven White
26-Sep-2012 13:16:40
An essay on the GNU web site (http://www.gnu.org/licenses/why-not-lgpl.html) makes a bit clearer the motivation behind the GPL. It's not just that they want people to have the freedom to understand their software. They also want to encourage the writing of free software. So this essay all but says that if you want to link to a free library, they want it to be a GPL library so that you will have to make your own source code free also. I did not quite understand that subtlety before.

So I would have to put myself on the side of a license that allows people to write closed programs if they want to. I would think that would get the most people using REBOL, especially if they can make money doing it.

Shadwolf
26-Sep-2012 13:33:27
I always put my source codes in no licence do what you want with my work and so far noone made a thousand billion dollars on my work so paranoid internet stop trolling :P

MIT seems great. But in the end the right thing to ask ourselves is do we want a futur for rebol yes or not? Actually rebol is pretty much dead and generate no money at all so or you let it die burry it once and for all, letting oblivion have the last word. Or you give it a true chance to live on its own life. This is the pass choosen and like many here I think rebol deserves more than a half assed licence that will bind people to publish only free software based on rebol.

shadwolf
26-Sep-2012 14:02:45
To tell the truth in the actual state of abandon of rebol even a MIT licence will not be the guarantee that rebol will have a futur. And that a community will emerge and be able to continue carl's work where it was left.

So my point is dead for dead why not giving at least a true chance. No one will ever be able to deny that rebol is carl's work and even less if he remains the official source of distribution for r3 and modules. Even if I do a commercial RUBOL based on REBOL and copying it 100% everyone will point their finger at me and I will have no clients.

Many were the companies doing that kind of move at the begining of LGPLed softwares with network monitoring tools for example and that lead their clients to turn their back on them and those companies went bye bye in no time.

shadwolf
26-Sep-2012 14:14:13
IT's more than probable that in the 2 next year even with a MIT licence r3 will be pretty much in the same state...

And I think Carl fear that overall that the free licence comes to late and that it changes nothing.

I think carl's remark "REBOL is not like other languages, and you will soon discover that REBOL source is not like other source." means that maintaining and upgrading rebol is a true effort it can't be an half hearted thing that as no true commitement of its participants.

And I don't see the futur dev community organising around the blah blah useless closed box that altme has become. Even r3 chat will mean that the participants of the conversations are a selected few... this is already in contradiction with the whole actual intent.

And it's not like alternatives to altme to exchange freely opinions and points of view about rebol (and not the bible...) doesn't exists but no one see the interrest in using them or organise things.

Being organised and planning things is the best way to get things progressing on a regular and coherent pace.

Bo
26-Sep-2012 16:09:29
Like shadwolf, I have also never published any of my work under a license of any type. Use it as you will, at your peril! Again, nobody has made a significant amount of money as far as I know with any of my code (except for me, that is). My code provides a platform for services for which I charge, but I don't charge for the software itself (not yet, anyway). It would be cool to know where snippets of my code are being used, like Code39.r (barcode generator), for example, but I obviously can live without that knowledge.
Bo
26-Sep-2012 16:22:13
I recall Rebol's source code from when I worked at Rebol Tech, and especially Carl's code was the definition of elegant. Not only his code, but also his design behind the code. I'm sure that's what Carl's intent was when he said "...you will soon discover that REBOL source is not like other source."

Almost anyone can write source, but it's another thing entirely to write source that not only works, but is elegant, efficient and has a solid design behind it. I naturally tend toward the former group (as probably most coders do), but I strive to be part of the latter group.

It's my opinion that any Rebol source contributors who want to have their code accepted as part of Carl's "official" Rebol release need not only desire to be part of the latter group, but demonstrate that they deserve to belong there through the quality of their source code contribution.

Bo
26-Sep-2012 16:31:48
I remember fondly the first thing that Carl asked me to do when I started at Rebol Tech was write a script. I submitted the script to him, which he looked over briefly and said, "Now make it 70% smaller".

I pulled all the tricks out of my hat to reduce it by 30%. Carl gave me a quick lesson, and I was able to reduce it by 70%.

My advice when you see the Rebol source code is to not dive into changing and adding things right off the bat. Study the source, look at the way Carl makes things happen, and learn from his vast wealth of experience and skill. Then, do your best to emulate the same level of skill when you add your mods.

shadwolf
27-Sep-2012 4:56:53
Bo what is proposed here is massively cooperative way of work. If you filter the access to what people will say or apport then you already cut the goal to have the most participant possible proposing things.

I said proposing things. I can propose a bunch of code lines to set a new fonctionnality or solve a bug but like you said Bo it will not be the best way possible. But the fact that I propose my bunch of code allows other to have a reflexion basis and then discuss on the best implementation possible which is the way we should work.

But by narrowing the access of the media then you make difficult the proposing and talks process.

My expericence tells me that if you say to RMA that their project is doomed to be a dead end since their is a black box they don't have control upon and that they are not very proactive in a their project management bam you get banned. If you profess jesus love on all discussion groups then people laugh at you but let you be...

What show that RMA are weak is that they never intended to convince their enemies to join them they pushed away all the people they didn't acknoledge or that had a negative feeling about their task. Those same people will probably be around and participate to r3 developpement (no fear if they produce so little as they did in RMA/GUI I remember you that rebgui was done by a single individual in less than 3 month)

Shadwolf
27-Sep-2012 5:07:09
Bo like you said not everyone as a vision and if you compare RMA/GUI and rebgui this is cristal clear.

So if I was in charge of RMA/GUI how I would have done things?

I would have tend to include all the people the most passionate by the VID/GUI. I would have propose to port and enhance rebgui to r3. I would have propose to add new things like svg renderer widgets or text rendering widgets. I would have said the people that don't or can't code can still do paper works, share their experiences and not only in altme a world that only 50 people read even if there is a web transcription I doubt that thousand of people a day read it. Since I'm banned I try to follow discussions that way it was a pain. And anyway reading interesting discussions without being able to participate is in any way possible frustrating and not the right way to handle things.

shadwolf
27-Sep-2012 5:10:09
Why should I possess a replica rolex since I can go to a pawn shop and have the real deal for a fair price?
zzz
27-Sep-2012 6:30:53
"Although it is unrealistic to assume we can filter/fix all impurities that might come about in an open environment, we should all strive to minimize the degradation..."

Would love to see the released src!

Arnold
27-Sep-2012 6:55:04
The same reason as you would want a replica REBOL, it is even less expensive and you can take it apart, see how it works and even add your own functionality to it while it still fits into the box.

Irritating watches and other spam. This morning I went again thru the effort of reporting them all..

Shadwolf
27-Sep-2012 9:10:01
Arnold, Tss, Tss... I don't want cheap things that fall apart or don't do the task they are supposed to handle. (most of the rebol replica don't have a parse to start, it's like a rolex replica without the clockwise ... :)
shadwolf
27-Sep-2012 11:14:20
for licence opinion you can still vote on http://gopollgo.com/what-licence-do-you-want-for-rebol-opensource

(no limit of time and there is some neat statistics like the country of the voters which I hope will please CArl to see that all over the planet this is a hot topic)

Steven White
27-Sep-2012 11:40:17
Looking through these messages, I see references to licenses that I never have heard of, and I suspect there are many more I never have heard of. I would wonder how many of those voting for MIT/BSD do so not because they think MIT/BSD is best, but because they have not heard of others that might be better. I am assuming that Carl has heard of these others, and is picking something that matches his goals and everyone's desires, which seem to be to keep REBOL going, keep it as close as possible to his vision, allow others to help develop it, allow people to write closed-source programs with it if they want to. I have not voted because I feel like I don't really understand the subtleties of the various licenses.
Arnold
27-Sep-2012 11:53:17
Shadwolf, I prefer a real Seiko over a fake or real rolex any-time ;) I am not that brand minded. But I still eat Raider instead of Twix and Treets and Bonito's not M&M's. In the case of Red 'versus' REBOL, I see Red bringing additional value to the REBOL language. And REBOL, well I hope it can survive and be a lighting example and guide in keeping the language compact and non-complex and certainly will keep being useful in many area's. So yes to both and no to bloatware-langs.
Arthur
27-Sep-2012 17:22:02
ahh, Thanks Carl. A difficult decision indeed, I could feel it in your writing. Now its like opening the sacred mathematical scripts of Egypt. We were blind but soon we shall see. I hope one day you bless us more with parting your design knowledge in some form of video series. The world needs it.

Thanks again, much love to you and the family.

David Cope
27-Sep-2012 22:53:06
Hello Carl,

Great to hear this news, and to read your previous blog post. Wonderful that you've resurfaced and to hear you are enjoying your work & life.

Having used REBOL in my professional work, I like to think I could do the same again. I'm no expert in licenses, but I'd like that we have one that allows REBOL developed apps/systems to be sold without source where appropriate, but also do the whole free thing too.

Could the future REBOL "inner circle" and we the general community consider a mechanism so we can request and track features and library development in a coherent way?

Need to revise my old REBOL skills now!

Brian Deneen
28-Sep-2012 0:08:30
Carl,

KEEP IT CLOSED! Please see the usual communication channel for my reasons. Hope to see you at AmiWest 2012.

Leke
28-Sep-2012 4:11:58
Awesome, it will be interesting to see how things will progress now.
Alexei
28-Sep-2012 9:21:49
When you release your project under GPL, you basicaly give up on controlling it. Some developers regreted going GPL and closed subsequent releases (for example Menuet OS). You can not expect people would understand and follow REBOL principles, though I may be wrong... In fact it all depends on if existing REBOL community is strong enough to keep REBOL "pure".
Peter R. Edridge
28-Sep-2012 11:21:14
Congratulations, I think it's an excellent move. I for one support the REBOL design philosophy, so colo[u]r me volunteering. I even write documentation without a whimper. :)

Greg Brondo
28-Sep-2012 11:52:03
Carl, Thanks so much for this gift! I look forward to the release!
Shadwolf
28-Sep-2012 12:04:20
"keep it simple" isn't quite the whole thing behind rebol. it's more "keep it simple without loosing efficiency"

Red is appart in the clone world since it openly says it is inspired by rebol and not tend to mimic it. When I was refering to fake rebol that don't have clockwise my mind was more set on freebel, borron, orca, ghost etc..

Gin
28-Sep-2012 14:03:34
Im also against GPL. Linux has 1% max of PC desktops because of GPL. GPL introduces fear and takes away your freedom. None of serious businesses will risk using GPLed software.

REBOL needs adoption, not GPL.

shadwolf
28-Sep-2012 16:15:41
how Carl will decide the licence for the opensourced rebol 3?

simple here is my suggestion:

rebol[]

index: random length? licences: [ "lgpl" "GPL" "apache" "BSD" "CUSTOM RT" "MIT" ]

print rejoin [ "Rebol picked this licence:" licences/:index "!!" "^M So now STFU because rebol said so !" ]

Only two lines of pure love to solve an epic problem!

did you smoke something
29-Sep-2012 6:14:35
(at)brian deneen

REBOL has already been "closed" for decades, and that's the main reason why it got to this moribund state today. The world has changed completely, it's not the 1980 amiga days anymore, the new generation only knows about Open source, internet, Windows, IOS, Android, MacOS, linux today. Don't think Carl can afford the same mistakes again by keeping it closed, if he wants his baby to have a chance of staying alive.

shadwolf
29-Sep-2012 10:26:11
(at) did you smoke something -> someone had to advocate for the other side. After all he represent a community trying to reborn a dead OS since 1995 or something like that...

It's a sink or swim situation here. Unlike amiga and amiga OS r3 have only one guy behind, so if the guy behind decide to let it down then it's dead and pretty much it's what we see actually. It's not that people aren't interested anymore in r3 or rebol in general it's that they don't have the means to keep going on... and no rebol will not by miracle survive if Carl don't enhance it, in it's r3 alpha stage.

shadwolf
29-Sep-2012 10:33:41
brian deneen -> you misunderstand what is at stake here I think... If Carl took that move it's because he sincerely think that he will not have the time to finish r3. He is fighting alone on 2 fronts the first one is correcting bug (which isn't a pleasefull journey in wonderland) and the other front is creating new technologies into r3 that will bring interrest to it and make r2 look so old that noone will look at r2 anymore.

The opening motion was already set brian, Carl choosed to put his trust in a selected few that pretty much told him that r3 was so bugged that it was impossible for them to do their parts(RMA/GUI). As the selected small openning to a select few of elites didn't worked so now it's time to open widely.

r3 had no new version since feb 2011 ... I know brian that for you 3 or 10 years without news is like a common thing totally normal.

Arnold
30-Sep-2012 4:19:37
Shadwolf,
REBOL forgot to seed
random/seed now

How can there be a time to harvest then
;-)

pepa
30-Sep-2012 6:34:21
REBOL is awesome! Thank you for this decision to open up the source, what a gift!! Huge GPL-fan here, but for a language framework such as this, an LGPL-license is more appropriate. Contributions to and modification of the existing code are to be open, but new code linking to it doesn't have to be. rBSD/MIT will mean you'll get loads of non-open projects based on REBOL.
test_user
30-Sep-2012 16:30:49
I'm waiting for the exciting time!
Chris Travers
30-Sep-2012 23:43:24
A couple points in reply to Maxim's points above.

The limits of the GPL, any version, are not well known and depend possibly to some extent on your jurisdiction. The GPL quite explicitly does *not* apply where no copyright infringement would otherwise be found, such as where fair use etc might apply. Moreover, it isn't clear how courts will apply the question of where the line between breach of contract and copyright infringement is regarding the GPL. Finally the GPL v3 is a horrible mess and likely would be incompatible with the BSD license if both are read literally (yet every open-source-involved lawyer I have talked to says these are compatible but for various different reasons which are themselves incompatible with eachother). I say all this as a developer, and owner of an open source software development firm, not a lawyer. The choice is then between a couple of options and models.

On the first, there is the idea of setting REBOL free, under a permissive license, like the BSD or MIT licenses. In this model Rebol Technologies becomes an expertise center, and sells various services relating to the software. Forks will happen, even proprietary ones, but chances are that if you can get the community together and behind you, the software will flourish well.

On the second there is the dual license model, namely being a single vendor solution, open sourcing R3, and selling proprietary licenses to anyone who wants to redistribute Rebol with their proprietary applications. Such licensees would get the sorts of protections Maxim and others want.

I think both of these will be extremely hard to directly monetize in the current market, btw. Programming languages have become effectively commoditized to the point where it is very hard to make money off of them unless you have a major name in the business, like Microsoft does (and even there, how long before MinGW becomes a major thing on Windows?). If it were up to me I would go with the permissive license model because it is likely to build the broadest possible community and hence the largest possible market for Rebol Technologies' services. However this has to be weighed against current revenue streams and it isn't clear what will be necessary short-term outside the inner workings of Rebol's board rooms and accounting offices.

Anyway best wishes to the Rebol team!

Arnold
1-Oct-2012 0:49:01
It is the first of October for the server too. Unexpected positive effects of spam discovered!

Today could be the big day!

I am curious what it is going to be.

Brian Hawley
1-Oct-2012 1:21:06
OK, let's break down what we can and can't do with this license, with help from some AltME discussions. First, the basics.

These licenses are based around a few concepts, such as linking and calling programs. It would help for us to look at what these mean for R3:

  • Static linking: For REBOL code, this usually means preprocessors like prebol or Ladislav's include. For R3 we add embedded modules and mezzanines, since the source of these are compiled in as string constants. Embedded extensions are also statically linked, as are the natives, more or less.
  • Dynamic linking: For REBOL code, this is what bind does, it dynamically links the words in your code to other contexts. You can also dynamically link REBOL code to extension native code using make command!, and native code links to other native code using the usual methods.
  • Calling programs: GPL-like licenses generally allow you to use other code as long as you go through clearly defined interfaces that are supposed to logically separate the code into separate programs, or something like them. Some platforms extend that to their plugin APIs, but some don't. For R3 code the only clear ways to make this separation are to use the call or browse functions, or sending data back and forth through ports. However, you could argue that using do or import on scripts or modules that only have external effect, no lasting internal state changes, is effectively the same thing, but the scripta are still dynamically linking to the runtime library. Anything beyond that would have to be an allowed exception.

Just in case we end up with a mixed license situation, let's break up R3 into functional categories (based on the current system model, for the new people here):

  • Natives and other cross-platform native code. In the current R3 host kit this is the stuff in r3lib.dll.
  • Host code, which is platform-specific native code that wraps the cross-platform code.
  • Headers and API specs, like the stuff you'd need to use to make an extension.
  • Mezzanine code, which is code written in REBOL that is built into REBOL, and used to implement REBOL.

It would also help to go over some of the standard things we would expect to be able to do with R3, that some of us do already:

  • Calling scripts/modules/extensions using do, import, do-needs, or from the command line.
  • Making extensions, either for speed, special tricks, or to wrap native libraries.
  • Making host kit apps, or porting the host kit to other platforms. These aren't necessarily programs themselves, you can also make hosts to integrate with other platforms or applications.
  • Learning REBOL by reading REBOL source, often by using the source function.
  • Participating in REBOL community projects.

Pardon the lists. They're not done yet.

Brian Hawley
1-Oct-2012 1:22:34
Let's get to the meat of it.

If everything is licensed as GPL2, here's how the above activities would be affected:

  • Every script/module/extension that you run would be subject to the GPL because it binds to the GPL'd natives and mezzanines. This wouldn't extend to scripts/modules/extensions that don't use any built-in natives or mezzanines, as difficult as that would be to do, but...
  • Even if you don't import an extension into R3, you would still be bound by the GPL if you build it with the API headers. And if someone wanted to port the API headers to Red or something, it's still considered a derivative work. That means that you can't wrap native APIs unless they have a GPL-compatible license, or are included in the operating system; other libraries aren't exempt in GPL2, not til GPL3.
  • Your host kit projects would also be bound by the GPL, with the same restrictions on the use of non-OS libraries. Also, you won't be able to make hosts that are plugins for closed-source applications.
  • If you make scripts based on what you learn by using the source function, you run the risk of it being considered a deriviative work, and thus also subject to the GPL.
  • You'll only be able to participate in REBOL community projects if you haven't seen the R3 source, or only in GPL'd projects.

That seems a bit harsh, but you might be able to negotiate or buy another license to solve some of these issues. Let's look at some related license models to see how they affect things.

Everything LGPL2: Since all external scripts and such are dynamically linked, they can be any license that isn't directly incompatible with the LGPL when linked dynamically, and that's not much of a restriction. Extensions can even wrap third-party libraries (as long as they're dynamically linked), but the extensions themselves would be subject to the LGPL because you can't dynamically link a header. Embedded modules and extensions would be subject to the LGPL because of the static linking. Everything else would still apply, except you could now participate in LGPL projects as well if you see the R3 source.

Everything Classpath: Like LGPL, except you would be able to make closed-source extensions and embedded stuff without needing to dynamically link. Still iffy to learn from the source and participate in community projects, unless they're at least Classpath-compatible.

GPL 2 with a special exception (like the GCC model): The stuff in r3.exe and r3lib.dll would be GPL, but there would be an addendum to the license that gives permission for user scripts to be run by R3 without making them subject to the GPL. Contributions to R3 would only be accepted if they agree to the same exception. If you mix in regular GPL code in your own builds without the exception, scripts that run on such builds would be subject to the GPL.

All permissive (like Python, Ruby, most other modern monolithic languages): No restrictions on use, or on what you compile it with, and you can learn and participate without limits. Not as much opportunity for RT to control/monetize REBOL directly, so he'll have to sell services or something.

Mixed license model (like Mono and most other modern complex development platforms): Have the natives and other stuff in r3lib.dll be LGPL, the mezzanines, host and API code be BSD, MIT or Apache 2, and r3.exe be GPL with the exception to let it run REBOL scripts. Find the right balance that will encourage adoption and participation, but still have the kind of control that will keep the design great. It can be done.

Does this make sense? Which should Carl and RT choose?

Luke
1-Oct-2012 2:12:49
My reading is that if REBOL was a serious earner for RT, justifying commercial investment and maintenance Carl would not be in this position of going open source. So monetising REBOL cannot be the primary consideration, but rather getting uptake and community input is. So if you want widest adoption I say go for a permissive license, like almost every other open language.
Nick
1-Oct-2012 4:52:21
Carl,

The posts by Chris, Brian, and Luke above are really worth reading. There has been an enormous amount of serious discussion about licensing in AltME, and Brian summarized it succinctly. From what I can see, the comments by Chris and Luke seem to reflect a general consensus among your potential community. My hopeful vote is that you'll choose the most permissive license, to build community and ensure progress, then monetize by selling services and support.

Luca
1-Oct-2012 4:57:59
Am disappointed with GPL license; it seems very "complexity polluted" to me.
Fork
1-Oct-2012 9:22:24
Those who have stuck on threads this long aren't statistically likely to be passionate about open methods for ideological reasons. Most would have been satisfied with a closed Rebol that could be bought for a flat rate like a chip (if the commercial product had been developed and shipped at a satisfactory rate, and the price was "reasonable")...

But copyleft helps prevent a situation like the current one from re-arising: a tool you enjoy that's held back from you being able to hack on its source. This helps whether you're tired of waiting on a feature, whether you want to audit for security, or whether you want a modification that a closed-source developer disagrees with on matters of "design aesthetic".

Thus LGPL has a danger—say something cool comes along and replaces VID. But this adaptation winds up not being open source because it just "links" to the interpreter. Again: people here would probably tolerate such a license on a product if it were actually good and released frequently. (Though they doubt that would happen, and that any closed effort in tool-building that starts in this era will not experience momentum or uptake.)

All that said, Stallman wrote of the LGPL:

"When a free library's features are readily available for proprietary software through other alternative libraries... the library cannot give free software any particular advantage, so it is better to use the Lesser GPL for that library"

It's hard to say that JavaScript or Lua are in the same league as Rebol, but they would be seen as tough competitors in the embedded scripting language space. So there's a balance of the importance of that vs. Lober(TM)(c)(R) coming along and being the new node of "this is cool and I want to use it but they won't give out the source and it's been 2 years since the last update".

onetom
1-Oct-2012 10:12:23
+1 for BSD/MIT

If anyone makes serious money from Rebol, I'm pretty sure they will pay more than just respect to Carl.

It's also not the best use of anyone's life-time to think about funky licenses so much, but instead think about and build software which harnesses computing power to actually help and enable people to create even more awesome stuff :)

It's already 2nd of October here, btw. I'm still holding my breath... ;)

Carl, can you share the visitor stats for these articles in comparision to the former ones, please? :) I'm so curious about what is the real interest in Rebol...

Brian Hawley
1-Oct-2012 10:26:54
Of course that assumes that proprietary software is unethical to create, which I will on occasion agree with. But I have to admit that I write proprietary software for a living, so if I can't write proprietary software in a language then I can't use the language, period. Most of the best REBOL developers are in the same situation, Kaj being a notable exception.

Even if I could buy a license, I would be disallowed from hacking on copyleft source because it would constrain what kinds of projects I could work on, even my job prospects. I don't sign NDAs for the same reason. For me, copyleft = closed, but with more staying power (we hope). This even affects Kaj, since working with copyleft R3 source would constrain his ability to work on similar permissive-licensed projects.

If R3 were GPL-with-exception, Classpath or (to a lesser extent) LGPL, but had MIT-licensed mezzanine code, then I would be able to use and contribute to the project just as much as I did before (time permitting). I could even do the same kind of black-box testing on the native code that I did before. If the host code were also permissive-licensed, even if r3lib is LGPL, we could do all sorts of porting and integrating without usage constraints, as it was for 2.100.111 and earlier. It would even be legal to make Visual Studio extensions, as an example. No loss of functionality.

If R3 went permissive then it could take advantage of the benefits of cooperating with other community projects. They would be complementary rather than competitive. And I could hack on the native source as well, as could Maxim, the Saphiron guys, Doc, and a lot of other REBOL gurus. Plus there would be no limits on adoption of the language, no legal drawbacks when compared to languages like Javascript, Lua, Python, Ruby, Go or Red. REBOL would be better able to compete with the permissive-licensed development tools that are becoming the new norm.

We have to pick our tradeoffs.

Shadwolf
1-Oct-2012 12:20:58
GPL ground was to promote open source and free software base on free software for free software. And to create a clear separation betwin commercial that looks like freeware (shareware) and true free software.

In order to maintain and avocate for the real free software world it was at that time mandatory to make a true distinction. We are almost 20 years after so obviously the free culture is more known and less restrictive licence emmerged.

Now is it normal that apple does billions of dollars on a BSD base modified OS? So would say "Hell no !! totally unfair !" other would say " yes it is because apple greately contribute to the BSD community financing it and giving drivers bug fix or event pushing major companies to produce driver for BSD-like OS."

This is why it's called an ECOsystem. Some will feed on others some will disapear and be forgoten, Some other will evolve and be better.

All we can expect is that this chance given to rebol to be reborn will have a concrete use and that efforts will be gathered into one single intent.

We should not forget that forks exists for several reasons, change in licence/owner ships, change in technologies, change in dev team, Impossibility for some talented individual to impose their views on the main stream. Fork isn't evil, fork is just another word for evolution process.

Those statements are based on concrete examples. unity fork from gnome3, cinnamon fork from gnome3, libre office fork from open office. and still there is project so big and so apealling that never forked, mono, blender, moonlight, linux kernel (tons of version still 1 single design, kernel and main dev), kde, QT, opengl, perl, python, gcc, etc, etc,

It's in fact pretty rare that a fork gets more attention that the software from which it's originated.

When talking about something in general it is better to have the whole picture. focus on the best aspect and try to apply them to your own situation to aim at success. Some people doing freesoftware don't aim at nothing else than knowledge so the success factor is de facto out of the scope of the discussion.

And yes at a pure knowledge stand point my personal belief is that carl code for r3 will inspire future generation for decades. That the least we can expect form the open source of r3.

shadwolf
1-Oct-2012 12:26:36
one example of truely usefull fork I would die to see (and that I will probably never see in this century :P) is a fork that base rebol upon java and see how it evolves.

It would be terrific in a close futur to see nullish things like the javaFX to evolve to better fit the needs and originalities of r3/GUI.

Nick
1-Oct-2012 12:30:32
Carl,

REBOL needs guys like Brian, Maxim, Saphiron, Doc, etc. Brian wrote above:

"If R3 went permissive then it could take advantage of the benefits of cooperating with other community projects... And I could hack on the native source as well, as could Maxim, the Saphiron guys, Doc, and a lot of other REBOL gurus. Plus there would be no limits on adoption of the language"

Aren't those the basic reasons for choosing open source at this point? Money can still be made selling support, services, products, etc. built with REBOL. A permissive license will build the community.

shadwolf
1-Oct-2012 12:32:20
(at) Brian-> hehehe, hum Richard Stallman would say this to you: "If you can't affort to get an ethical job in software programming then change job. There is less need of software developper than on cops, road builders, teachers etc..."

And by the way most of the freesoftware devs are university professors (including RMS, which actual main activity is to travel the world to diffuse the moral on software realisation)

Yes RMS is a bit too much of an idealist for he's own good. But knowing this allows you to understand why the GPL is the way it is. GPL and FSF want to stand as the more ethical idealistic thing possible in order to pave the way for the futur generations...

shadwolf
1-Oct-2012 12:39:01
clearly on the poll I ran it appears that a majority wants a MIT/BSD like licence. (few people voted but you still can vote http://gopollgo.com/what-licence-do-you-want-for-rebol-opensource)

And once again it is not because you have access to the source code that you will understand it and be able to modify it.

It's not automatic I can give you hundred of open source project that will take you decades to fully understand.

Example? ok go read Jokosher source code then try to add your little something to it that will spice it and then try to convince the main stream dev team that your little something is a true plus that makes the things better and that is a must have for the next release :P

Omei
1-Oct-2012 13:06:39
Assuming if i look at source i am taintet: If i ever looked into system/words, i am tainted too. WHo of the gurus can prove they did not?
Brian Hawley
1-Oct-2012 13:52:56
Shad, I definitely appreciate the humor in this. I have an ethical job in software programming already. There are many different ethical principles to follow and some of them conflict. I can't follow all of the FSF principles at my job without violating federal laws, some of which are good laws, and for that matter endangering our customers' personal information, which for me is a greater ethical breach. This isn't the only field for which this is the case. RMS can be a bit narrow-minded at times.
wt
1-Oct-2012 20:13:01
what's the time now? ... only 4hours rest.
Arnold
1-Oct-2012 21:21:36
getting used to this?
test_user
1-Oct-2012 22:10:35
I believe that Carl will keep his promise!
Bo
1-Oct-2012 22:44:37
Carl said: "My schedule is to finish this up next weekend and make the release by October 1st."

Forgive me if I sound like I'm coming to Carl's defense. I'm sure Carl is doing everything he can to get this released, but as we are (almost) all programmers, you'd be an unusual one if you always met all your schedules.

I'm waiting on pins and needles as well for R3 to go open source, but I'm also realistic in understanding that 7 days may not have been enough to make it happen, especially when Carl's days are taken by other responsibilities.

Carl has made the decision to go open source, and it is progressing. The waiting is the hardest part for the rest of us.

"Patience, Grasshopper."

Carl Read
2-Oct-2012 1:32:29
I note it is still the 1st in Hawaii...
voxel
2-Oct-2012 8:11:16
Carl you're the best !

Thank you a zillion times :-))))

Ingo
2-Oct-2012 13:33:03
Hi Carl,

first, thank you again for creating REBOL.

I know it is your "baby", but as you obviously stopped caring about it, I believe the best you can do is: give it the most open license, and give the community a chance to build itself around it.

It would be great if you could build it, but as you aren't there to do it, this is the next best thing.

Ingo
2-Oct-2012 13:37:27
(at)Bo

It's not about letting the deadline pass (again). We all know that.

It's about letting it slip "without notice", and the disregard for REBOL and what is left of the community

Arnold
2-Oct-2012 14:03:01
Not being able to get things delivered on time happens all of the time. Even me :) but I get away with it because I communicate this as soon as I have a clue that it might occur. Being the boss does not release you from letting know your customers, personnel etc what is going on. But I am in no way capable of judging what is happening at RT HQ. A short notice to the community saying sorry it is taking me longer because real life events take more time or anything else would have been kind.
Omei
2-Oct-2012 14:23:15
Carl comes with a generous offer, community says: Waaah!! You want to taint me!! Maybe he is unhappy somehow?
Fork
2-Oct-2012 14:33:57
To be fair:

(a) Prepping code for publication (scrubbing any private data, cleaning up or deleting any comments that might make one look bad, removing temp code, properly establishing project structure with Git and GitHub) can take longer than one thinks.

(b) The fervent debate over license here (with "I wouldn't use it, it would be dead to me" reactions from some, as opposed to a chorus of "sounds fine") is reasonable cause for delay from schedule, due to further consideration of the issue.

Knowing that the world is checking in for developments...an explanation would be ideal. But if one only does blog updates and does not "tweet", the idea of an entry and ensuing comment flurry being necessary for a self-imposed deadline might seem like "noise". I can somewhat empathize wanting the next entry to be "here it is, done"...and a few days margin of error on an estimate being acceptable.

While we wait and see, you can help me shorten my bingo program :)

http://codegolf.stackexchange.com/a/8558/57

Bo
2-Oct-2012 16:49:03
(at)Fork...Uh, the only way I can see to shorten your bingo program is to take out the comments! You're insane.
Brian Hawley
2-Oct-2012 21:43:06
Ladislav has reminded me that the stuff I've been saying about linking restrictions and such depend on the GPL being enforceable in your jurisdiction, or at all. Consult a lawyer before doubting that though.

Now, back to the bingo!

shadwolf
3-Oct-2012 4:59:31
Brian unfortunatly there is no humor in the quote I did from RMS in fact it is a quote from a resent interview he gave for the episode 200 of linux action show by jupiter broadcasting.

http://www.youtube.com/watch?v=radmjL5OIaA

It's educating to see that interview from the main Guru itself. (by the way the goal of this interview was initially to discuss means to make money and ear their living for freesoftware developpers)

ok so where are the r3 sources and what licence are they?

shadwolf
3-Oct-2012 5:19:14
hum my analyse on the actual talks about the licence is not as drastical than it appears.

Everyone else than Brian Deen rejoice for the rebol opensourcing.

But there is a BUT some very implicated people in this community want to work for rebol developement at the same time than working for red developement. They want to feed rebol with the new things they learnt from red and feed red from things they could learn from rebol.

For example red do an extensive use of rebol parse, so it is logical to imagine that parse will evolve and be enhanced to fit the best the red use of parse. This is a mutual benefit no? it's red needs giving purpose to rebol's features. With a restrictive licence like GPL v2 you will not be able to set that kind of motion.

And it is not about being tainted or negative stuff like that it is not even about not recognising the wonderfull gift Carl is about to do!

shadwolf
3-Oct-2012 6:03:26
Can we say that Carl is not right to fear that red will salvage the dead body of rebol for its own benefit while letting rebol3 die and decompose?

So if r3 was a human being lets say he had got involved in an ugly car accident and he is in a deep coma state brain dead and the doctor comes to the familly to ask them if they will donate his organes to a poor little child in need name red that only chance of survival is to get those organes.

We can see the MIT/BSD demand from some people in the community that way. But I know those people I know that they will not do that kind of move and that they will try to make red and rebol benefits from each others.

GPL says that all the work based on it or derivated from it (recuperating parts of it) have to be free. but it is Free this doesn't means it can't be sold ... it means you will broadcast widely the content of your source code and share the information about it. Free here stands for the possibility for the client to see how the software is made and being able to modify the code the way they want even reselling it after their own modification until they share their source code and openly recognise that they base their work on previous work from previous authors.

MIT/BSD licence lead to close source code for part of your solution. So in the case of red is it normal to suddently see appear a "black box" bit code that contains codes copy pasted from r3 would that be acceptable?

Freesoftware GPL based is not equivalent to anti commercial !

Free software means the user the final client can look as much as he wants to the source code of the software he use.

shadwolf
3-Oct-2012 6:15:45
Firefox is a freesoftware, not only because you don't pay to get it but mainly because you have access to the entire content of it's source code. Tons of work are based on (ice weasle, etc) it and the only restriction that exists is that those work sold or not diffuse the full content of their own source code and recognize that parts of those source codes are copy-paste and customization of another software.

So you can't suddently do your windwolf browser sell it diffusing only the binary and deprivate the end user from the freedom to know how your software is done and if there is spyware/adware included in your binary software.

Don't forget that commercial software includes auto deactivating routines if you don't buy the next year licence. So you end constantly pay for a service.

shadwolf
3-Oct-2012 6:28:11
actual licence of red based on MIT/BSD is logical since red include the need of a commercial product no source code shared named rebol. Since red need that commercial closed source binary piece of software to work then you have to have a licence that can fit the both need.

The need for red to diffuse it's source code and be share with or without direct or indirect charges. And the need to being tied to a black box commercial called rebol (which actual isn't sold but still relies on a commercial closed and restrictive licence).

This is why MIT/BSD is needed in red.

If tomorow I want to do an software in red and sell and that rebol and red and GPL all I will have to do is in about page or documentation says that my work is based on red and that the source code of my own software is available on my site and that my sotware's licence is GPL.

Once again it is not because you people can read my source code modifies it and resell it that you will do it. You want a concrete example of that?

Lets take viva-rebol! or area-tc. Noone retook it for it's own benefits. Same with MDP-GUI etc... Or cheyenne! or mysql-protocol. Noone even retook freebel to enhance it and make it's own rebol-clone based on a java engine.

See we return to the things I was saying earlier it is not because you have access to the source code of something that you will change it.

shadwolf
3-Oct-2012 6:54:38
Lets continue the reflexion of a commercial software based on GPL and difusing widely it's source code because he is too in licence GPL.

The logical ask is how will I make my work profitable since everyone can access my source code modifies it or not and sell it. Well first thing I will say that my work is the main stream that I have the background knowledge to maintain and improve it which is not the case of my competitors that only resell my own work. I can say too that my competitors only harvest my own work for their own benefits without returning the favor to me in any way. Or I can simply constantly improve my own work to make the competitors looks not quite the thing. I can do like redhat enterprise do say the licence paid is the opening to high quality support service. Etc...

Red hat is a good example there is a hundred plus linux distro but red hat still does more than a billion dollars of benefits based on the reselling of a gnu/linux GPLed OS. So when it comes to say that you can't make money with something based on GPL this isn't quite the truth. But yes the main question is What commercial strategy you will apply to your specific personnal solution to make your living and for that there is no magical solutions.

I will remind you that red hat enterprise has actually two non commercial derivating one originated from redhat company the famous fedora and the other being made by a totally independant association named CentOS.

REdhat is sold fedora and centos are not sold and independent version of the same thing and still redhat does a billion dollars in earnings every years. And every years it increase.

So when people says I can't do money on a GPL based software I say ooooooh really? but would your software make a money in a commercial based licence anyway?.

Like rebol, everyone likes it, everyone acknoledge Carl's talent and efforts he put in it everyone qualifies rebol as a true need for computing area but still rebol as product is a commercial faillure. This shows that licence and money making aren't tied. Firefox made a tons of money more than rebol ever did and even the most funniest thing is 80% of that money is based on free will donation from none of the less than MICROSOFT !

shadwolf
3-Oct-2012 7:28:17
more than the licence for opensource what really REALLY interrest me is what will happend next to rebol.

Ok the licence everyone wants is adopted and rebol is opensourced including rebol 2.

What is the next move. How things will organise? Who will be the main branch ? Will the forks main interesting points will be polished and retro integrated into rebol main branch or will the rebol main branch dev team will be so lazy that naturally the eventual forks gain more power more features and more interrest than the main branch leading to the main branch disapearance?

What are the plan for rebol main branch in the next year? how many release a year will be afforted?

How will be financed the everyday rebol leaving (website maintainance for example etc..) ?

I'm affraid that the main questions are completly occulted by the licence concern which is in reallity irrelevant.

If red people don't want to participate in rebol futur that is their own concerns.

If they don't want to have code copy pasted directly from rebol I would say that is for the best, because red doesn't need to be seen as scavanging a dead corps for it's own benefits.

and at the same time I really feel it would be a big loss to have those two project not benefiting from eachone and having more than a different goal in betwin.

I want connections betwin red and rebol I don't want red to be forced to drain the human ressources from rebol in order to exist which would lead to rebol disapearance because noone will serriously organise the work around it or plan for improvements.

Sure this should be Carl's main task but do Carl want or have the time to do that?

shadwolf
3-Oct-2012 7:33:26
As for the miracle hypotetical unknow mass that will suddently appears and take rebol organise things integrate it in a major product like a webbrowser or a image creation tool.

I'm sorry, but this is an eventuality that I don't see happening. If those people existed they would had already massively manifested their interest in rebol's opensource.

But who knows?

wotWok
3-Oct-2012 9:04:29
So excited, thank you so much Carl!

And to the future core dev team, please keep REBOL under 1 MB always, cheers! :)

shadwolf
3-Oct-2012 9:10:09
Maybe tomorrow when Carl release the full opensource rebol 3 source code, we will see a massive "Hurray!" from the computing industry. All the main actors will gather and found an organisation and allocate resources to make rebol the world unified script language for everything with an os and a CPU!

Yes can happend I strong doubt it will but it is still a possibility.

What makes me doubt that this will ever happend is that no much people shown their interrest in any clone of rebol to enhance them including red.

I think it will anyway take time for a global motion to be set and that people will be attracted only if the rebol project has a previous rock stable way to work and produce release on a regular pace. A clear direction to the rebol futur is mandatory too. No plans no appeal. Go to blender.org look how they work how they advocate for the next version content and how they deal to raise funds for the next version. CAuse since 1998 there is no new version of blender that had not been previously financed. When they write lines of codes in blender that is because they always received money to do it!

Carl Sassenrath
3-Oct-2012 22:21:19
This is sure a busy group of people. So many comments.

It looks like there's some concern over GPL v2?

I work a lot with GPL v2 code every day and its use in commercial enterprise... that even gets blended with other licenses. It does not seem to hinder decent product development, although some care must be taken. I cannot say the same about GPL v3 which is a big problem, and I avoid it.

Also, I figure that GPL v2 is perhaps the most tested and mature of all the models, and we know that very large software communities endorse and support it -- such as Linux.

However, that said, let me go back and read more of your comments to understand your main points. Sorry about the delay, but this part I've got to get right. What licenses are used by other popular languages? Ruby, Python, Perl, Java, Javascript, etc.?

Arnold
3-Oct-2012 23:03:51
Thank you very much for your reply, Carl!

Completely agreed that you need more time to do it right, you only got about one shot ;)

So take the time, and please "keep us in the loop".

I am pretty sure a list of languages and licenses will show up within the next few comments.

Regards

Gregg Irwin
3-Oct-2012 23:04:43
http://en.wikipedia.org/wiki/List_of_open-source_programming_languages

Lua = MIT Perl = Artistic/GPL PHP = BSD Python = PSF/GPL Ruby = Ruby/GPL Dart = BSD Scala = BSD

Gregg Irwin
3-Oct-2012 23:10:56
From the Ruby license:

5. The scripts and library files supplied as input to or produced as output from the software do not automatically fall under the copyright of the software, but belong to whomever generated them, and may be sold commercially, and may be aggregated with this software.

A note about the PSF (Python):

Unlike the GPL the Python license is not a copyleft license, and allows modifications to the source code, as well as the construction of derivative works, without making the code open-source.

Gregg Irwin
3-Oct-2012 23:19:55
The main thing for me is the "some care must be taken" part. My clients generally don't have legal departments to review this stuff, so it would fall to me not to put their IP at risk. Full disclosure about possible risks might deter them from letting me use REBOL. Almost all the people I work with support the giving back of generic code that they didn't spend a lot of money to develop, so that isn't an issue.
Graham
3-Oct-2012 23:32:30
The main issue is that some people interpret the GPL to mean that any script aggregated with the R3 interpreter eg. as in encapping, would make that script a derivative work of R3. The other view is that the script is just data for the R3 interpreter, and therefore not a derivative work. This may not have been tested in a court of law yet.

The simplest thing to do to settle the concerns is to specifically state that scripts written, and aggregated with an unmodified R3 binary, are not considered a derivative work by RT.

Otherwise, as the copyright holder, you may well be contacted by recipients of 3rd party encapped code asking you to enforce the license and make that 3rd party provide the sources of their scripts.

Graham
3-Oct-2012 23:56:08
Ruby changed in 2010 from a dual GPLv2 + Ruby license to BSDL + Ruby license as Matz felt the GPLv2 was inhibiting the writing of free software.

Ref: http://www.ruby-forum.com/topic/216010

AdrianS
4-Oct-2012 0:17:06
As Gregg said, the problem with GPL2 is that you do seem to have pretty decent legal help to be able to feel at ease using it. This kind of experience is lacking in many companies. In my experience, the interpretation being relatively ambiguous (without deeper analysis) has made any company I've worked at take a pass on using any libraries using this license. Please, just keep it simple - why not MIT or BSDL?
MaxV
4-Oct-2012 0:51:15
The owner of Rebol is Mr. Carl Sassenrath and Reboltech. If they will not start a lawsuit to who use Rebol, where is the problem?
Rebol license could have a preamble that explain that: intepreter and script are different, even encapping them; intepreter is under GPL, your script not.

Arnold
4-Oct-2012 1:05:15
MaxV. There may be not a problem now, and not coming from Carl, but if there is money involved you never know what sharks may come after you when you have made some money using REBOL to do so. You might even get sued for far far more than you ever were able to generate.
Pierre
4-Oct-2012 1:16:11
I voted MIT/BSD on the poll. But, finally I think that maybe a GPL* with the appropriate addendums could be just as fine as a MIT/BSD, no?

I took a look at the licences of the pythons running on my linux box: frightening! Python 2.6.6 licence shows 280 lines, and python 3.1.3 licence has 284 lines of boring terms... The diff is almost nothing, only in history.

If anyone is interested, I could paste here the licence; but, frankly, I think all of us are quite able to type licence() at any python prompt...

I looked a bit closer: in fact, python (which has a development policy which seems to me quite similar to what Carl imagines: a main coach, a main team, and contributors) (which is very successful also) licence is not a GPL licence, but it is GPL-compatible (well, most of the versions).

(1) GPL-compatible doesn't mean that we're distributing Python under the GPL. All Python licenses, unlike the GPL, let you distribute a modified version without making your changes open source. The GPL-compatible licenses make it possible to combine Python with other software that is released under the GPL; the others don't.

I agree with you, Shadwolf, that the main issues do not concern the licencing, but, rather, the real future of Rebol and its various derivatives, notably Red.

Although it is important to release Rebol under a proper licence, which should not have any negative side-effects, maybe writing a "KISS" licence, rebol-style, very straightforward would be faster than choosing among

Carl, thanks for your feedback: please take your time, give yourself a more realistic deadline (please keep us informed!).

Cheers!

PS: I hate reading or writing licences. I rather prefer writing code. And reading code, also... And running code, most of all!

Carl Read
4-Oct-2012 1:41:10
Carl, it might help the discussion if you clarify what you/RT want the license to achieve and to allow.
Luis.
4-Oct-2012 1:47:42
It seems fairly obvious that one type of licence generates confusion, concern and doubt. The other doesn't.

Either choose the one that doesn't or generate your own.

I thought you wanted to keep it simple?

Cheers,

Luis.

WuJian
4-Oct-2012 2:07:17
Simple description of popular software licenses

http://paulmillr.com/posts/simple-description-of-popular-software-licenses/

I vote BSD or MIT(even better),let REBOL stand on the same starting line with other big languages, and let's see REBOL run! Hello, world!

REBOL can bring much to the world,and the world will pay much back to Carl.

Best regards.

GrassRoot
4-Oct-2012 3:16:14
I wonder if the wrapper to make an encapsulated exe is included in this open sourcing ...
ecollage
4-Oct-2012 3:41:45
Yippee yai kay! Thank you Carl!
lui
4-Oct-2012 4:06:34
License Chooser

http://three.org/openart/license_chooser/

Nick
4-Oct-2012 5:15:40
The link by WuJian makes clear the limitations imposed by each license. A huge problem, however, like Gregg and others have pointed out, is PERCERPTION. Just as many potential supporters have immediately rejected REBOL in the past because it wasn't open source, many of the most serious developers will again immediately reject the open source project the instant they see a restrictive license.

*** We've already seen this reaction from some of the most important and supportive developers in our own community. *** That should hit you like a ton of bricks.

My hope is for REBOL to grow as large a community as possible, so that it can thrive. Any sort of restrictive license (i.e., not MIT), is a potential road block to REBOL's acceptance. I think choice to open source REBOL is a fantastic move, and you are at a turning point in finding support among talented developers around the world who will build and use REBOL, keep it current, support it, and improve it. If you want to see REBOL become as successful as possible, it's in your best interest to remove any restrictions from the license, and grow the community.

Steve White
4-Oct-2012 9:14:18
I am staying out of the license discussion because I am totally unqualified to have an opinion. But I check this page every day am glad to hear that progress is being made because I work as an in-house programmer for a non-profit organization, and the topic of a programming tool keeps coming up. I am lobbying for REBOL, but concern was raised about that plan when it appeared that REBOL was near death. Now that it appears that death is not imminent, the discussion changes.

I also would like to second those who thank Carl and ask to be kept "in the loop." A comment now and then that things still are moving along is reassuring.

Scot Sutherland
4-Oct-2012 9:22:52
Disappointed but not surprised.

The license is critical. I will need to control my IP written in REBOL because most likely the result will be a product not a service. Service providers can use open source more freely because they are paid for their expertise and time, not the product.

My dissertation work relies heavily on REBOL 2. I would like to move to R3 in the future, but will stick with REBOL 2 if it allows me to keep my IP. Or I may have to leave REBOL, which would greatly complicate my situation.

Scot Sutherland
4-Oct-2012 9:37:36
Also critical to my work. It is very important that I am not forced to reconsider my code whenever an "update" occurs. Everything I have written in REBOL since 1.3 continues to work. I appreciated that Carl decided to start R3 rather than making updates to R2 that would break older code.

Strict adherence to the principles will help. For me REBOL has reached it's 18th birthday. Just like people, REBOL will begin to die. How gradually it dies and the sort of impact it will make depends in large part on how well it is cared for. Poor choices will shorten its lifespan.

Hopefully REBOL will produce new offspring that are not open source. We shall see.

Carl, everything I know about programming that really matters has come from looking at your code, so I will enjoy seeing how you started with code that to machines and designed a language around people.

My deepest thanks. Can't wait to see what you do next.

Brian Hawley
4-Oct-2012 9:51:21
What licenses are used by other popular languages? Ruby, Python, Perl, Java, Javascript, etc.?

It depends on the implementation, since you can't actually copyright the language itself. Most of the older languages that predate the OSI invented their own licenses, but more recent languages tend to reuse existing licenses because there's a backlash over license proliferation that's been going on for the last decade. Even some older languages are changing to standard licenses now.

Ruby: Some have the Ruby License, which is a custom license that has an explicit fallback to the 2-clause BSD. The old official Ruby interpreter dual-licensed the Ruby License and GPL2. The new official interpreter was Ruby License/GPL for one version, briefly GPL2 only for 2 minor point versions, then switched to straight BSD after some community feedback (similar to the comments here). JRuby is CPL/GPL/LGPL to be compatible with OpenJDK. IronRuby is MS-PL. MacRuby is Ruby License. Other modern Ruby implementations are BSD.

Python: CPython and JPython have the Python Software Foundation License, which is a BSD-style custom license that is compatible with the GPL and OSI and FSF certified. IronPython is MS-PL. PyPy is MIT.

Perl: Perl 5 uses its own Artistic License 1.0, dual licensed as GPL (don't know which version). AL1 is not technically an open source or free software license, though they originally intended that it would be, which is why most projects that use it are dual-licensed with the GPL. Perl 6 uses its own Artistic License 2.0, which is a copyleft license with different terms than the GPL, but which has a relicensing clause that makes GPL-compatibility possible. For the AL you have to provide the source of any modified versions that you distribute back to the original copyright holder, along with a description of the changes, and give permission to the original copyright holder to use your changes in the original product. It's also copyleft to other people you distribute it to, and you aren't allowed to charge money for your changes, and since the original has accepted changes under this license without copyright assignment they can't charge for that either. You can charge for distribution though. You can link to anything you like, even closed source stuff. Commercial-hostile to make, commercial-friendly to use.

Java: Oracle JRE is closed source, though it's based on OpenJDK which is currently GPL with the Classpath linking exception. They have a compatibility testing kit (TCK) which has a open-source-hostile license that pretty much prohibits competing open source or free software implementations. Thus, Apache Harmony can't be said to be compatible with Java even if it is, and that's Apache 2 licensed. If you don't license one of the official implementations, you don't get the patent licenses. Oracle barely gets away with this because people are used to it being evil, and the language is entrenched so many have no choice.

Javascript: Most implementations are built into web browsers, and thus are licensed like applications that are meant to be extensible with arbitrary-licensed code - commercial, MPL/GPL/LGPL, LGPL, BSD. V8, the most popular out-of-browser implementation, is BSD licensed.

Brian Hawley
4-Oct-2012 9:54:38
As for languages/implementations introduced in the last 5-10 years, most of the ones with significant uptake or backing are are Apache 2 or MS-PL if they have patent grants, BSD or MIT if they don't. This includes Lua (MIT), Scala (BSD), F# (Apache), Go (BSD for the main implementation, and the other primary implementation is built into GCC and is licensed like that), Vala (LGPL because it builds on glib), and recent entries like Dart (BSD), Rust (MIT), CoffeeScript (MIT), and TypeScript (Apache). Microsoft started using MS-PL, but all of its recent releases have been Apache 2.

For big platforms, it depends on their aim.

GCC is designed to be hostile to closed source derivatives and extensions, but friendly to closed-source use, so it's GPL with an exception that lets you link to all libraries in it as long as they're either compiled only with GCC or only without GCC, basically. Copyleft but not use-restricted, wide entrenched use.

LLVM is designed to be friendly to all uses, so it and its libraries are provided under the NCSA license, which is like a cross between the 2-clause BSD and MIT licenses in a way that is compatible with both. Very permissive, and very widely used, more all the time.

Mono is designed to combine code from a variety of sources with the LGPL'd code of the GNOME project, so it has code under a variety of licenses, from LGPL for the stuff that links to glib or Qt, to MIT for the stuff in the Mono libraries proper, to Apache and MS-PL for the stuff from Microsoft, to GPL for a few of the top-level applications, to commercial closed source for parts of the Android and iOS ports. No copyright assignment either so it will stay that way. Very friendly to commercial use, and very widely used.

Most other platforms in common use are commercial with some permissive-licensed libraries, and are backed by Microsoft or a major hardware maker. There are also some commercial tools that are entrenched in financial, industrial or government circles, though government is increasingly going to permissive licenses (when contractors develop it; government employee stuff is public domain).

Fork
4-Oct-2012 10:05:09
Thank you for the update, Carl.

I will reiterate/agree that starting from what the license is designed to enable or prohibit is the main point. Legalese follows the intent--and there's lots of choices for implementation of intent (including writing one's own license from scratch if one wants, though those are rough waters.)

Fact: For tools used in software development, people do not want to have the tool dictate the license for the thing they are building. This is nothing new; pick a language out of a hat—Microsoft Visual Basic, for example—and it has no such restriction. I would think this uncontroversial. But given the spread of reactions here, perhaps there needs to be a from-the-horse's-mouth statement of this...maybe laid down in the license terms somehow.

When I look at the situation I put myself in Carl's shoes, and I'd feel a quid-pro-quo factor. If someone finished R3 and built the next generation web server called "Node.Rebol" I'd be enthusiastic...but a lot less enthusiastic if they weren't open. It's not about money, so much as about profiting from a derivative while the community doesn't get anything back.

I don't see why the Mezzanine couldn't be MIT/BSD, if that pleases BrianH. I still don't understand why that matters, but what's the harm? (I thought it already was a do-whatever license from some prior conversation.)

Can people deal with GPLv2 for interpreter, MIT/BSD for Mezzanine...and an explicit statement that you can license programs built on the foundation however you want? LGPLv2 would please more people and perhaps permit some dodgy embedding, but honestly probably won't matter in the near term at all.

Again: if the main trunk keeps a contributor license agreement clause, all this can change if things aren't working out. It seems like a more-than-more-than-fair starting point when giving out a whole lot of work. I feel all this resistance and bickering is just delaying us having a chance to tinker, and see what we're even talking about!

Brian Hawley
4-Oct-2012 10:19:07
A couple widely used exceptions with different licenses than the above:

The Boost C++ libraries use their own Boost Software License (BSL), which is a permissive license that's less restrictive than the MIT, 2-clause BSD, or NCSA.

SQLite is explicitly public domain, and you have to give up your copyright and put your code into the public domain to contribute to the project or they won't accept your stuff. It's extremely free to use, but might be iffy in countries that don't have the public domain as a concept in their copyright laws. I haven't heard of that being an issue yet.

Brian Hawley
4-Oct-2012 10:28:29
Fork, my contributions to R3's and R2's mezzanines are MIT licensed. If they need to be something more permissive like BSL or WTFPL let me know and it'll probably be OK :)
droid rebelion
4-Oct-2012 11:43:25
Kivy seems to be okay with using GNU LGPL Version 3.

(Windows Linix OSX Android)

http://kivy.org/docs/

apotheon
4-Oct-2012 11:44:23
GPL has been a dominant license for a while, but that time is waning. These days, people are increasingly releasing things under copyfree licenses (such as 2-clause Simplifed BSD License, MIT/X11 License, Open Works License, TESLA, ISC License, et cetera) instead of copyleft licenses (GPL, LGPL, MPL, et cetera), showing that the growing wisdom of collective experience has found the supposed benefits of copyleft licensing do not measure up to the hype, and the detriments are real and worth considering.

Commonly overlooked problems of copyleft licensing include mutual incompatibility between different copyleft licenses, growth industry in reimplementations of copyleft projects under other licenses thus fragmenting the "market" to its detriment (not all fragmentations are detrimental, but these are), reduced commercial usage (and thus reduced commercial contribution to the core project), and the necessity of managing copyright assignment for contributions so that future legal matters are manageable by the project maintainer(s). Even sharing of GPLed software can open one to legal liability in some cases thanks to the requirements for long-term maintenance of source access for the specific version of the last binary distribution, without recourse to linking to an upstream provider. Further, the actual legal tests the GPL has passed are essentially limited to decisions that an open source license has recourse to the law, rather than necessarily being toothless just because money may not have been made off the license terms by the open source project maintainer.

The most common excuse for using the GPL is a desire to "force" those who use it with modifications to contribute back to the community. The realistic result is that those who would otherwise use it might just not use it. If they did use it under a copyfree (rather than copyleft) license, they still might contribute later, and even if not they improve the popularity of the language itself (in the case of a programming language implementation like REBOL), which helps improve the popularity of the reference implementation by providing wider portability and compatibility of source code, in much the same way that commercial implementations of the OpenPGP protocol help improve the popularity of open source implementations by providing wider portability and compatibility of the protocol. For many reasons, some languages' implementations are changing their licenses toward the copyfree, e.g. Ruby's reference implementation and Python's PyPy. Copyfree licensed Clang is rapidly claiming copyleft licensed GCC market share, too.

Apart from all of the above, the matter of whether to require copyright assignment for contributions has no clear correct answer, but the answer must be gotten right at the very beginning before all facts are obvious, because there's almost never any going back on that decision. On one hand, you might end up driving people away from contributing by asking them to assign copyright to you, which might restrict adoption at a critical point in community building; on the other hand, without copyright assignment you will almost certainly not have the opportunity to later impose copyright assignment short of rolling back the whole project, and thus will not have the sole copyright control necessary to make decisions critical to the health of the project. A copyfree license solves this problem neatly, because such licenses impose no need to require copyright assignment while still allowing for reasonable control over project direction (rather than having to rely on the unreasonably difficult task of getting all contributors to agree to any policy proposals).

The only "good" reason to insist on copyleft licensing (with copyright assignment) is, ultimately, to abuse the community via commercial dual-licensing schemes that provide paid-only licensed enhanced versions of the software on the backs of community contributors.

Shadwolf
4-Oct-2012 14:46:03
I will make a movie called the Licence war out of this thread!

Jokes apart it's interesting to see that there is as much free licence than linux distributions.

I don't see why MIT/BSD would not be adopted. we all know that the day Carl wants to return to closed commercial licence for rebol he will do a rebol 4 from scratch that will makes look rebol3 like a useless toy. He has that talent an noone can take that from him.

Giuseppe Chillemi
4-Oct-2012 15:31:38
Carl, what you want to obtain?

R3 embeddable in commercial software or not ?

Modifications to r3 opens sourced or not ?

Modules open sourced or not ?

Extensions open sourced or not ?

Scripts encapped with r3 i think you want them to allow closed source.

Answer these questions and you will find the licence

Brian Hawley
4-Oct-2012 19:23:10
From Andreas in AltME:
Are you sure, Brian? Afaik:
  • Ruby 1.9.1 and 1.9.2 where also Ruby License/GPL dual-licensed.
  • Ruby 1.9.3 is Ruby License/2-clause BSD dual-licensed.
A quick look into the respective release archives seems to confirm that.

I'll take your word for it, though someone needs to fix the Wikipedia page. I remember the community discussion that led to dropping the GPL, though only second-hand through my pro Ruby friends.

It's weird, the Ruby License itself explicitly dual-licenses as 2-clause BSD, so why would they need to dual-license it a second time? Legal stuff is strange.

Fork
4-Oct-2012 20:07:39
(at) apotheon I'd never heard of "copyfree"...but read over your website advocating the term/concept. Hmmm.

I watched a recent interview with Richard Stallman. He described how his programming career began in the 70's, and how he'd had the opportunity to glimpse a "free" and sharing software community was. Then commercial interests came along and made it disappear; he wanted to bring it back.

(Today I showed a friend a "code wheel" that used to ship in video games, and she was baffled such a thing would have existed. I was a teenage hacker and NOP'd the JSR that called the subroutine that did the check...or inverted the condition so the only way it wouldn't let you play was if you gave the correct answer. :P)

Yes, cultural tides have shifted. Inert code capital is about as profitable as a dead-tree book, and everyone is making money on services/control/invasion-of-privacy. One might argue that the GPL "did its job" in changing norms, and the battles of the EFF are more relevant than the FSF. It's telling that the GPL has incorporated Affero clauses for server-side programs, v.3 revisions to stop "Tivoization"...trying to stay relevant as the code diminishes as a piece of the profit puzzle. :-/

-but- your claim that the GPL is "waning" is simply not true. Someone ran a script on recent Debian releases and found that the proportion of packages licensed under GPL-family licenses since the release of GPLv3 is increasing, not decreasing. Sure...search GitHub or wherever shoddy JavaScript codebases are found in spades, you might think otherwise. But there are countless copyleft success stories...not just Linux & its packages, but Wikipedia and StackOverflow.

Rebol's tagline is a "rebellion against software complexity". That is an ideology--or at least it sounds like one to me. People who are practical and wanted to make lots of money have avoided Rebol like the plague. Less brilliant folks than most commentators here fired up text editors and frenetically programmed in languages like PHP (which is truly "A Fractal of Bad Design")...and made millions/billions. They're still doing this!

But just as Rebol shouldn't give in to peer pressure to be crappy and expedient for "practical" reasons, I don't think it's fair to throw the good ideas of the GPL overboard for "practical" reasons. The license-mixing argument frustrates me, because it's really the other licenses that are the problem. In an all-GPL world everyone could have what they wanted.

Arguing otherwise is like throwing your most reasonable and intelligent friend out of a party, because noisy ignorant jerks don't like them. You should throw out the noisy ignorant jerks and stand with the person who talks sense. And I don't want to buy a "cool" Rebol-powered gadget that locks me down, steals my personal info, and I can't customize.

To me, Carl is driven by the "philosopy-of-complexity-management"...and I'm a believer. But I also believe in the "philosophy-of-sharing-alike-for-software-freedom". Having the two meet and make a stand is nice in my book. I think that this also sets up a natural divide: the pragmatists go with Red (which is doing rather well), the religious people go with Rebol.

Additionally: the people who seem to think Carl has the time or interest to come and sue you for looking at GPL'd code...you've probably never been in court. It ain't fun for anyone; costs time and money. If Carl doesn't have time for Rebol, he certainly doesn't have time or interest to sue you over it...unless you're Apple/Microsoft/Google/etc. I sincerely doubt anyone here is thusly affiliated (we have ONE person from IBM? And IBM sells *Linux* clusters anyway!!! GPLv2 is not bad, people.)

-pekr-
4-Oct-2012 21:09:06
Fork, from the legal perspective, it really does not matter why you or I or anyone else thinks about Carl eventually not having time to sue anyone. There is either chance to be sued, or not, and if the risk is there, ppl will opt out from using such solution, easy as that.

For me, it is OK to be pushed to publish any kind of changes/improvements etc. to the REBOL itself, but it is not imo OK, to be pushed to release anyone's app IP, completly unrelated to REBOL. And GPL pushes you to release sources to all your app code, doesn't it? If so, it is total crap, which, if choosen, will burry the REBOL chances for many, the second it happens ...

Lee
4-Oct-2012 21:41:30
That is a great news. I don't like GPL too. Or you just mean only the source of REBOL is GPLed, but the user can use his owen license for his application; if so, there must be a good REBOL compiler/binder to hide the source
Marco
4-Oct-2012 23:14:48
Hi Carl,

Stop listening the community !

Many years ago, you announced for Christmas a new version of Rebol with enhenced GUI. After that, you start listenning Rebol's user and the announced GUI was never released.

Same with this opensource release, if you continue to listen the community, you will never release it !

Regards, Marco.

Carl Sassenrath
4-Oct-2012 23:16:39
Thanks for the insightful comments. I see that I need to be more clear. Let me explain what I want to achieve, and let's discuss it and settle on a license that works best toward that end. (See next article.)
apotheon
5-Oct-2012 23:32:03
Fork:

It takes a strange perspective to look at the world of open source software, see all the myriad of different licenses out there, notice that the GPL doesn't play nicely with any of them (including licenses that are essentially exactly the same as the GPL in effect but bear different names and use slightly different phrasing) -- to the extent at minimum that GPLed code generally can't be incorporated into differently licensed projects, and often can't be combined with code under other licenses at all such as in the case of something like the CDDL and other GPL-like copyleft licenses -- and decide that the problem is not the GPL. Oh, no, it's not the one license that can't play well with others that is to blame, in your view. The real problem is EVERYONE ELSE IN THE WORLD.

I really don't understand that perspective at all. I find it simply baffling.

Bo
16-Oct-2012 12:02:54
OpenCOBOL celebrates Rebol going open-source:

http://www.opencobol.org/modules/newbb/viewtopic.php?viewmode=flat&order=DESC&topic_id=1499&forum=1&move=next&topic_time=1348456309

GIT-GUI
6-Nov-2012 11:35:33
as stated elsewhere slax has restarted and Slax 7 release candidate 1 is out but not got the standard Development tools module as yet so no easy way to make a new REBOL 3 module for it yet and try and get to it included before final, but then theres still no R3 source to build a slax module ether...

http://www.tomas-m.com/blog/19139-Slax-7-release-candidate-1.html

https://github.com/rebol/r3-source 1 file

LateCummer
23-Aug-2013 5:27:19
Too little, too late. REBOL doesn't even make it into the TIOBE Top 50 programming languages list. Might have been a hipster language when it debuted, but now other script languages ( i.e. Python ) are now the defacto standard.

Post a Comment:

You can post a comment here. Keep it on-topic.

Name:

Blog id:

CS-0511


Comment:


 Note: HTML tags allowed for: b i u li ol ul font p br pre tt blockquote
 
 

This is a technical blog related to the above topic. We reserve the right to remove comments that are off-topic, irrelevant links, advertisements, spams, personal attacks, politics, religion, etc.

Updated 21-Oct-2017   -   Copyright Carl Sassenrath   -   WWW.REBOL.COM   -   Edit   -   Blogger Source Code