Mathias:

Just to clarify on one point,

>somebody can write such a long blog and still misses the important details I explained here.

I don't know who you refer to by 'somebody' here, but in case it's referring to me, the reason I didn't include what you explained here is simply because I simply didn't know the discussions that took place between the ESC members. I am not on ESC and I do not routinely follow what goes on on that mailing list. What I wrote on my blog is my personal experience (after all, it's my personal blog), and I can't write stuff I don't even have knowledge of or I have not experienced.

Kohei

Posted by Kohei Yoshida on October 06, 2007 at 10:41 PM CEST #

I always try to stick to facts and based on the facts that are known to me it's as I wrote: I'm totally baffled and have no explanation for that.
The reason is that I couldn't believe that Michael Meeks (who is in the ESC as the representative from Novell) has not informed you about these things as they are very important for you and your work. Michael always complains about bad communication in and from other teams (there is a long track record for this in the ESC minutes) and so I must assume that he will inform a member of his own team about such important things as soon as possible. At least in no longer time than a few weeks. Is that in inappropriate assumption? I think not.
So perhaps you should have a talk with him why he withholds such important information from you and at the same time blames others for treating you so badly.

So if you didn't know all this it makes the whole story even more sad.

Posted by Mathias Bauer on October 06, 2007 at 11:25 PM CEST #

Mathias:

But judging by what the ESC mailing list archive contains, the only thing I've missed is that you guys were willing to accept the Solver as an add-on? I didn't see anything else that is substantial that I could've missed. What you detailed above seems like a long version of why the Solver cannot be accepted as an external component (which Michael did inform me, of course).

Besides this specific issue, what I went through as a contributor which I outlined in detail on my blog is real. There is no denying that. And as I said at the beginning of my blog entry, I was not trying to put a blame on anyone, but to simply highlight what I experienced as an outside contributor (before I joined Novell), and how we need to change the current development process to make it easier for future contributors. And based on the feedback I received from others, others also share similar experience as well.

If you see my blog as an offense or insult, please accept my apologies. That was not my intent. But I believe my story still needed to be told in order to highlight the problem the project currently suffers, so that we can work on finding a resolution in order to attract more future contributors, and to make all parties involved in the project happier.

Kohei

Posted by Kohei Yoshida on October 07, 2007 at 12:12 AM CEST #

ok... so the solver can't be made into an extension... why, again?... I didn't get that :S

Posted by phobos on October 07, 2007 at 05:18 AM CEST #

Can somebody take this case of misapprehension and work to reverse this unfortunate turn of events? It is truly sad to see something that is akin to a fork rather than something like Symphony which revolves around collaboration. Novell's OOo edition for Windows (with all the VBA and 'extra' stuff) was enough of a burden because it made OOs incompatible with itself (assuming different platforms). It needn't get more severe an issue.

Posted by Roy Schestowitz on October 07, 2007 at 09:57 AM CEST #

I remember a story... about when I was young, a friend of mine who lived a few doors down the block had a wonderful sandbox. It was maybe 20 feet by 20 feet in size. For us little fellows it was huge. There was plenty of land space there for many forts and roads made by our little trucks, steam shovels, bull dozers, etc. Well, one day I was there and built a complex of building using sand and stone block (for roofs) that I was proud of. The quality was close to perfect. Guess what happened next? The boy whose parents built that sand box, he came over (was jealous for some reason) and executed a "right of ownership", he cut back on my "section" that I could work in and as part of this move, this resulted in the complex of roads being no longer possible at all between my buildings! I of course protested, and he then said this "it was his sand box". Well, it was clear to us all then (as there was 2 other boys playing), that this could happen again at anytime that we "built something better" (to anyone of us). The other boys were on my side during this disagreement. And I never went back to play in this sandbox ever again. And the other boys didn't play there either. We took our fork and played in another sandbox. And the sandbox went unused after this as the boy who owned it was left there to play alone, and he didn't really like to play alone I guess. I see much the same thing going on here. Sun needs to reexamine it's role with Ooo and to open up the sandbox to more democratic methods of both ownership and management of the code. If they don't then, they risk doing more harm then good, for what reason? Do the folks at SUN get a powerful "ownership" feeling that the JCA gives them, and really that is it's only value? Novell, as we have seen, is not pure when it comes to playing well with others in the community either. When you think of it... the GPLv3 is really a wonderful thing. Everyone again gets to play equally, and that is what the goal is anyway. Microsoft, who does not let anyone play in their sandbox without agreeing to terms that gives MS everything in return, is sitting back and loving this story. Sun and Novell need to wake up. Maybe they should all talk to RMS and see what he things is the best thing to do...! Really let him judge and then go do what he says is "right" and maybe then everyone will be happy.

Posted by Anonymous Reader on October 07, 2007 at 01:08 PM CEST #

Dear "anonymous reader",

you presented a funny story but it totally misses the point of my blog. We never wanted something threwn out of our "sandbox" because we felt we could do it better. This is exactly the point why I wrote my blog: the Novell campaign makes people think this because it mixes up two unrelated things. And you also fell into that trap. So let me try to make it clear again.

We didn't accept the "Kohei solver" only for one reason: because it is not provided under the JCA. And to keep your sandbox analogy: we don't want to have others bringing in their sand into the project's sandbox without getting the right to move the whole sandbox including all the sand brought in by various people to another place in case there would be a need to do so. The GPLv3 you mentioned would be such a case: without the JCA it would be close to impossible to make that move in a reasonable time frame (if at all).

The Calc team still would accept Kohei's code (and would be very thankful for it!) if he provided it under the JCA. It's not too late for this!

That the Calc team now will start to write its own solver component has two reasons:

- we can't accept the one from Kohei as explained above
- we have promised this component to our users for more than a year

We did the latter because until recently we have been under the assumption that Kohei's solver would be provided to the project under the rules of the project that have been known to and accepted by all involved people. This is what I wanted to explain in my blog. We don't want our users to suffer from the political struggles others have imposed on the project.

Especially the fact that we have announced Kohei's component as part of OOo3.0 since more than a year should convince you that we don't want him to go out of our sandbox. It's the opposite: he is very welcome to provide even more nice sand buildings.

Posted by Mathias Bauer on October 07, 2007 at 01:37 PM CEST #

As someone who had the unfortunate pleasure to deal with Sun in another open-source project, I can fully relate to Kohei. It is not only the JCA which is a PITA and requires one to hire a lawyer to get it translated into plain English, but it is also the attitude of the Sun employees. They ooze an incredible atmosphere of arrogance. Sun can never be wrong and all Sun is looking for are free-of-charge open-source programmers who do what Sun commands them to do. If you exhibit signs of own thinking you are the enemy.

I am actually very surprised that Kohei keeps so calm. Would anyone have pulled that Google Summer of Code stunt on me, I would have gone nuclear. And then the Calc Solver 3.0 announcement ... That's so typical Sun. Treating everyone else like shit. They like to do that. E.g. in the Java world they screwed Blackdown in a similar way.

I think Sun owns Kohei more than one apology.

When it comes to Sun-controlled open-source projects there is only one advice to give: *Don't join a Sun-controlled OSS project*

This includes:

- Don't sign the JCA. It is written in such a way that a layman can't figure out which rights are signed over to Sun.
- Don't discuss with Sun employees. They are by definition always right.
- Don't trust the word of a Sun employee. It isn't worth $0.02.

Novel is also not sitting at the table with clean hands. Their Microsoft deal smells. They certainly have good reasons to avoid agreements like the JCA like the plague. But the JCA is so bad that one really can't recommend that a developer signs it.

Posted by Hans Argo on October 07, 2007 at 01:42 PM CEST #

The (alledged) fact that the person was notified beforehand of the licensing situation doesn't excuse the point that contributing to a LGPL project requires you to give away your work to Sun under a closed license.

Think about that for a second, a person wants to work on a LGPL project and submit code. He can not since apparently sun also tries to sell the code he will write.

Thats not open source, thats unpaid labour.

And you know, OpenOffice can still benefit from this patch, its just Sun that can not.

Posted by T on October 07, 2007 at 02:18 PM CEST #

the contributors licenses are not uncommon in big open source projects and in the long run, they are there to protect the project...

Sun's SCA http://www.sun.com/software/opensource/contributor_agreement.jsp is not that different from Apache's ICLA http://www.apache.org/licenses/icla.txt or MySQL CLA http://forge.mysql.com/wiki/MySQL_Contributor_License_Agreement ...

you can read Kalle Alm's blog for an idea of why a CLA is "necessary today, where it previously was not" http://kallewoof.com/2006/03/24/contributor-license-agreements-and-the-effects-of-the-devil-aka-sco/

and I still don't know why can't it be made into an extension if the SCA is unacceptable...

Posted by phobos on October 07, 2007 at 02:47 PM CEST #

Hi "T",

I alleged nothing, I just drew conclusions from what is visible to me. Kohei meanwhile told that he wasn't aware of everything. OK, I never wanted to make this a "Kohei problem". He knew that the JCA was needed (he actually signed it), he obviously didn't know what other options he had. So there is a communication problem inside Novell and it makes the campaign they are riding even more dubious.

More important to me is that there are two different things we are talking about: the admittedly unfortunate history of the solver component and what happened to it after these problems got finally solved. Whatever happened to the solver after the denial of the JCA is a totally different thing that is not related in any way to anything that happened before. I know that some people have an interest to hide this from the outsiders view but only for their personal goals and crusades.

We don't do Kohei a favor in still mixing these things.

You also seem to misunderstand what the JCA is about. You don't give something away, you keep your copyright. And Sun does not "sell" the code, it is also using it of course. I don't see this as an unfair deal as more than 90% of the rest (without that the contribution wouldn't make sense anyway) is provided by Sun developers. I am not a lawyer (as I assume applies to most of us). But at least I wanted to correct your misapprehensions that are obvious even to people without knowledge about law and licencing.

I don't want to further discuss the licencing in my blog, people get too emotional about it even if they don't know all the necessary details. The stupid comment from Hans Argo is a good example that is not worth replying to except if one likes flaming - I don't. Only this: it's easy to call names on people one doesn't know but it's stupid and impertinent. All people knowing me will confirm that I'm not arrogant in any way as is true for all Sun employees working on OOo. We are not telling others what they shall do, on the contrary: other people approach us and ask us for help for what they want to do with OOo. Even if we failed to do that at times (and Kohei's solver unfortunately is an example for such failure) I would be astonished to see much current examples for this. A huge amount of our working time is invested in helping others to contribute.

Posted by Mathias Bauer on October 07, 2007 at 03:35 PM CEST #

I've read the Sun SCA, and see nothing scary in it, except the part on patents is a bit unclear and I had to re-read it several times. The agreement could be simpler, but isn't some big scary document. It is obvious to me that, the agreement states the owner retains ownership and also grant ownership rights to Sun, which in exchange promises to also release any changes under an OSI or FSF license. It's also plain that while Kohei signed an SCA prior to joining Novell, it would be necessary for his new employer to also, in respect to this project, grant an SCA on his behalf so that Novell (or someone else) can't come along later and say his contributions after such and such a date are owned by Novell (or someone else). Or Kohei would have to show that he retained his right to ownership of any code he writes while employed by Novell. I doubt Kohei can show this, nor most any employee of most any company.

However, if I ever work for another software company other than my own, I would require the company to sign an SCA with me so that I can retain my own knowledge and right to use it. After all I have written a great deal of source code both professionally and for fun, that has never been published. How could I protect my ownership rights if I couldn't prove I created the work prtior to joining a company. It is my belief that the current legal system does not protects the employee as it should. While I understand that any employee that comes up with useful knowledge (source code, patents, etc) while employed by a company, it is my belief that such a person should be able to retain the rights to use that knowledge in the future. The knowledge came about after all by the unique synaptic connections of that person and should belong to that person. This is the big problem with the current system and every employee should make their employer sign such an agreement. This would eliminate a great deal of lawsuits worldwide. Imagine if everyone's ideas belonged to themselves? It's a dangerous idea! It might be very damaging for patent trolls and abusive companies everywhere.

I don't know how Kohei was treated by the project, but that is irrelevant to whether the code can be incorporated in OOo, or as an extension. From what it appears here, it certainly can be made an extension, so what's the issue? Make it an extension; end of story. Personally, I think someone is behind all this bruhaha about open source projects being in a state of discord. A new FUD machine is here, and it's being very sneaky now. Must be a tactic from Sun Tsu's "Art of War".
Or a sign of growing pains. However, I believe the F/OSS community is a chameleon and will adapt to the new climate and survive and flourish. After all, a light on a hill cannot be hidden.

Posted by Brian Densmore on October 07, 2007 at 06:45 PM CEST #

Kohei,

I nearly missed your second comment. Of course it's not your fault that others used your case to do create noise and more over brought it into a context where it doesn't belong.

I agree with you that a lot of things went wrong in the past, in terms of cooperation and communication. And I agree that it is good do document what went wrong so that people can learn from it and do better in the future.
My problem is what was made out of your blog. What I mean becomes clearly visible from e.g. some of the comments in it. Especially the infamous comment from Fridrich (even topped by his own blog) and how Michael Meeks (mis)used your case for his crusade against the JCA makes me sad, angry and frustrated.

Again: the problems you had got nothing do with licencing. The licencing confusion started when Michael asked you to deny the JCA (as you wrote in your blog). I want to keep both things separated. Otherwise your worthwile attempt to improve the processes can become superseded by the "political" struggle.

Posted by Mathias Bauer on October 07, 2007 at 07:07 PM CEST #

I have tried to follow all the issues raised and I have learned a few things...

1. It does not look to me like anyone is completely blameless. These things sometimes happen and the real question is will each side come to the table willingly work things out?

2. Besides OOo Licensing issues just suck! Everyone has their own opinions on why they like or don't like a license. GPL2, LGPL, GPL3, MPL, BSD, etc. I know that Mozilla went on to try and "tri-license" much of the code to show openness.

3. A JCA/SCA type agreement may seem to "corporate" but if done right providing protections and options for both parties I see it as much of benefit as a hindrance. I have heard other unrelated groups comment that it might be be hard for Linux to move to GPL3 even if Linus wanted to. However I am sure that if Linus had a JCA/SCA type agreement from everyone he could move it to GPL3 (again another OSI License) if he wanted to do so.

4. If you don't like the ground rules setup at the beginning work to get them changed not circumvent the process. Remember when you live in your folks house you may have to follow their ground rules (Sun set the rules). However if you move out of your folks house don't expect them to keep coming to your rescue and funding your partying lifestyle and at that point everyone can lose.

5. I don't always agree with what Sun does, but if you want to do your own thing and think you can do it better then maybe Sun should just stop and let everyone else do it.

6. It is hard to keep up but it looks like OOo is missing some nice features besides Calc Solver because people wan to do their own thing. Am I reading right that if things had been worked out the following in OOo 2.3:
-Calc Solver
-VBA support
-Improved Startup performance
-GStreamer integration
-MS-Works import
-WordPerfect Graphics import
+more
Per http://go-oo.org/discover/

7. Didn't I heard Novell already had a Microsoft OOXML (MS Office 2007) Filter?

8. Some people hate "big business" and don't trust big business no matter what.

9. A few from Sun, a few from Novell, a few from the community, along with a 3rd party moderator all need to all sit down in a room (real or virtual) and politely work things out. Nobody will get everything, if things are not worked out then everyone will loose!!

Alan

Posted by Alan on October 07, 2007 at 09:31 PM CEST #

"4. If you don't like the ground rules setup at the beginning work to get them changed not circumvent the process."

You mean, do it Sun's way or be slandered in a Sun developer's sun.com blog?

With all these dark insinuations about Novell's motivations for forking OOo ... it needs to be pointed out that the fork happened in 2003. Ximian forked ooo-build precisely because Sun was talking bazaar but operating OOo as a cathedral.

So Novell-Microsoft conspiracy theories, and allusions to such, are not a complete explanation for the fork - that would require an explanation that holds for as far back as 2003. I'd be interested to hear the explanation.

Posted by David Gerard on October 07, 2007 at 10:24 PM CEST #

David,

Nobody should be slandering anyone.....

Just like in point 4 parents should not be slandering their children either....

I don't care when the fork happened ... we should always be careful of a fork. Not all forks are bad but it so splinters development ..... The license wonks understand what happened the developers try and keep up and the users give up as soon they get confused as "Bob" down the street has this version of OOo and Sally has this version OOo. All came out at about the same time but each has different features and different bugs .

Users (i.e. my Mother, my Boss, etc) don't understand forks, don't need 5 versions of a product, they just want it to work....

There may be a time with maturity and market penetration that more versions can be acceptable, but people need to be careful.

Posted by Alan on October 08, 2007 at 04:01 AM CEST #

David,

I don't think that I slandered anybody. And yes, if you don't like the rules you shouldn't try to circumvent them by pretending (or creating the impression) that these rules are the reason for everything that went wrong. In fact this was what happened with the solver component.

I didn't complain against the withdrawn contribution, I complained about the noise that accompanied that and that damages the whole project. If a contribution is withdrawn it's not OK to blame the project for looking for a replacement. If someone wanted to change the rules of the project he should not try to achieve that by using contributions as a weapon (or a hostage, depends on your viewing angle).

About ooo-build: you are missing an important detail. Whatever Novell had put into ooo-build from 2003 until mid of this year, there always was the will to upstream that eventually. So it wasn't what I would call a fork but a temporary "decentralized" branch of OOo. Now, where some of the contributions in ooo-build will not get upstream, it looks different. So whatever the reasons for ooo-build have been in 2003 now there is something else. It's not "dark insinuation", it's counting 1 and 1 together. And for me it's not surprising that the result is 2.

Posted by Mathias Bauer on October 08, 2007 at 10:29 AM CEST #

Is there any reason why Novell shouldn't properly fork OOo?

Please consider it, Mr Meeks. It has become profoundly obvious that Sun's instincts don't match those of our community, and as such, they cannot be trusted with one of our largest bodies of code.

Sun's OOo team are the new XFree86 Inc.

Posted by Martin Coxall on October 08, 2007 at 01:07 PM CEST #

If Sun would ditch the now utterly pointless StarOffice and release OpenOffice to be a proper open source project (ie maintained by a community, not a committee), the only thing that would ever matter is code quality.

You don't see Linus demanding people assign copyright to him and how many times has the kernel been sued.....ah yes, none.

Please get over yourselves Novell and Sun and just write some useful code so OOo can stop being the huge beast that it is and start rocking our worlds. We (by that I mean users - you know, your target market?) just don't care about your spats and legalese mumbojumbo. Your legal department might, but the appropriate choice of licence makes them entirely irrelevant.

Posted by Chris Jones on October 09, 2007 at 12:45 AM CEST #

Chris,
StarOffice is not pointless. There are good reasons for some users/customers to prefer it over OpenOffice.org. Google is a good example. But this is not important to this discussion about the JCA. Sun does not need the JCA for StarOffice. We could release it even without it by using LGPL only. We are not doing it now as it isn't necessary but we wouldn't have any problem with "LGPL only" for StarOffice in case of.

Posted by Mathias Bauer on October 10, 2007 at 11:05 AM CEST #

The two issues are related. The JCA is being used as a fulcrum to emphasize Sun's poor treatment of the OpenOffice community. If you stop treating them as unpaid serfs and start engaging them as a community, I predict the JCA issue will go away. Yes, you're saying Sun is going to improve in the future, but I'll believe it when I see it, as it's been 4 years or something that OO has been a "community" project. As it stands, OpenOffice is still community-unfriendly, so targeting features in to ooo-build is a great way to make you wake up and smell the coffee.

Did you see the comment by a founder of NeoOffice? http://slashdot.org/comments.pl?sid=316393&cid=20837491 "NeoOffice users have loved having the functionality ooo-build brings currently and continues to bring in the future, and much of the work pioneered by ooo-build is critical to maintaining the Mac platform as a viable office solution (read VBA). Sun's lack of acknowledgement and incorporation of ooo-build features does nothing but hurt users. Having received a "you're welcome to join us" response similar to Kohei, I am glad I do not consider myself part of OOo any longer. The freedom of forking has allowed NeoOffice to incorporate all good code without all of these politics and marketing games. Forking has allowed NeoOffice to deliver to Mac users the features they wanted yesterday regardless of where those features came from. Sun has a history of a "not invented here" syndrome at times when it comes to code within their "open" source projects."

Posted by James on October 15, 2007 at 05:58 AM CEST #