When I hear the word “fork”, I reach for my gun. OK. Maybe it is not that bad. But in the open source world, “fork” is a loaded term. It can, of course, be an expression of a basic open source freedom. But it can also represent “fighting words”. It is like the way we use the term “regime” for a government we don’t like, or “cult” for a religion we disapprove of. Calling something a “fork” is rarely intended as a compliment.
So I’ll avoid the term “fork” for the remainder of this post and instead talk about the legacy of one notable open source project, OpenOffice.org, which has over the last decade spawned numerous derivative products, some open source, some proprietary, some which fully coordinate with the main project, others which have diverged, some which have prospered and endured for many years, others which did not, some which tried to offer more than OpenOffice, and others which attempted, intentionally, to offer less, some which changed the core code and other which simply added extensions.
If one just read the headlines over the past month one would get the mistaken notion that LibreOffice was the first attempt to take the OpenOffice.org open source code and make a different product from it, or even a separate open source project. This is far from true. There have been many spin-off products/projects, including:
- StarOffice (with a history that goes back even further, pre-Sun, to StarDivision)
- Symphony
- EuroOffice
- RedOffice
- NeoOffice
- PlusOffice
- OxygenOffice
- PlusOffice
- Go-OO
- Portable OpenOffice
and, of course, LibreOffice. I’ve tracked down some dates of various releases of these projects and placed them on a time line above. You can click to see a larger version.
So before we ring the death knell for OpenOffice, let’s recognized the potency of this code base, in terms of its ability to spawn new projects. LibreOffice is the latest, but likely not the last example we will see. This is a market where “one size fits all” does not ring true. I’d expect to see different variations on these editors, just as there are different kinds of users, and different markets which use these kinds of tools. Whether you call it a “distribution” or a “fork”, I really don’t care. But I do believe that the only kind of open source project that does not spawn off additional projects like this is a dead project.
Odd, when I hear “fork” I reach for my knife.
But you can also see that the branching out is happening more in recent history which indicates a growing disappointment with the way Sun handled OO.o and latest Oracle.
What also is interesting is that these individual projects are of different sizes and different reasons. Just to call them all a fork or derivative without quantifying and qualifying is not giving the true picture of the situation.
Nevertheless i appreciate your data collection for the time line. That is indeed well done. May I challenge you to come up with a time line that indicates the quantification of each project and a separate one to show the purpose of the different projects.
Fundi
Years ago I remember getting some cheap VIA motherboards that came with a disc labeled “SupremeOffice”. From what I could tell it was a really bad re-branding of OpenOffice (1.x at the time). I can’t find any web presence for the project, I assume they died out at some point.
@Fundi, I don’t know if the increased number of OOo-derived projects signifies “growing disappointment”. Surely, we have more Linux distros today than 10 years ago. Does that also indicate “growing disappointment”? I don’t think so.
Some of these rebranded versions are more in the nature of a distro. They take the OOo “kernel” and repackage it with their own branding and extensions. In some cases they make core changes as well as extend. And in some cases they make more extensive core changes. That is one dimension of the differences.
The other dimension is whether the derived product makes upstream contributions to the OpenOffice.org project.
I think those two dimension could be illustrated, maybe with a quadrant diagram, but you would lose the time dimension.
I think the most telling discriminant that you could add to your chart would be whether the OOo “distro” is open sourced or not. I’m not familiar with all of the listed variants, but of the ones I do know, only OOo itself and GoOO are actually open source projects. Of course LibreOffice joins their ranks.
Anyway, the crux of the matter is that LibreOffice is intended to be a true fork, and independent of its parent project. LO will not be dictated to by Oracle, and will only accept code that takes LO in the direction that they want to go. And, the submissions to LO will not require copyright assignment, which pretty much guarantees that Oracle won’t touch them.
“Fork” is not a fighting word. It is a declaration of independence. There is only a fight if the forkee decides to fight.
Cheers!
Mark S.
@Mark, Unfortunately my chart can easily only illustrate historical reality, not “intentions”. Maybe someday LibreOffice will have a long string of successful releases and spawns off its own set of derived projects. If/when this happens, then the chart would obviously look different.
I’m trying not to get too much into the politics of this. I have friends on both sides of this issue. But it seems to me that the RedOffice and Symphony examples are interesting models. There is no doubt that these OpenOffice derivatives had added many substantial and distinguishing features in their products. For example, RedOffice has added many features needed by the Chinese domestic market, including special Asian text layout options, diagonal table headers, handwriting support, etc. And Symphony has added Eclipse integration and a new extensibility model. Both products do quite well in their markets. They both have a long history of releases, especially RedOffice, with a decade-long history of releases. But neither one of them has set up a competing open source project. Both of them take code from new OpenOffice.org releases and merge it into their own code bases. And both of them contribute some (but not all) enhancements back to the OpenOffice.org projects. Both of them publicly identify themselves as members of the OpenOffice project.
So, I’m not sure I see the radical “differentness” of LibreOffice. Aside from disavowing OpenOffice and all that political stuff, aren’t they just taking OpenOffice code and adding their own enhancements? Are they really doing anything at a technical level that would be worthy of a single paragraph in the press, if it had been done in Symphony or RedOffice? If so, I have not seen it. I think the noise is all political and the noise is much louder than any technical agenda or customer-facing vision that they have articulated at this point. OK. I can understand that. They are in the “community building” phase now. This is understandable. This will take some time. But at some point they will need to identify a customer/audience and argue what their value proposition is relative to other alternatives in the market. I look forward to hearing what that argument will be.
For me, fork does not necessarily mean dissatisfaction. It also means: let’s try something a bit (or a lot) different. Let’s experiment. “La richesse vient de la diversité” (now try Google translate on this you guys :-) or read about “Ethnic diversity” on wikipedia).
Let’s hope that all this creativity brings great new ideas / products!
Hi Rob, good post and interesting conversation here. If we can keep the tone as constructive as it is that’ll perhaps bode well for inter-project cooperation.
You raise the question ‘what makes LibreOffice different from all the other spin-offs of various sorts’. I think there are two main things.
Firstly is the way in which they started, they have arisen at a time when dissatisfaction with Sun and low confidence in Oracle converge with a booming number of Linux desktop users, or I guess GNU software users generally. So there are a lot of eyeballs expecting the LO project to raise the standard – that’s why they say they’re there. They have set the bar higher than other spin-offs and seem not to shy away from admitting that they are a true fork aimed squarely at taking the place of OOo.
Secondly I think the nature of the response they got seems to show that many OS vendors are quick to jump ship. So these OS vendors (like Canonical and Red Hat) now have a very public vested interest in actively supporting LibreOffice rather than OpenOffice.org. I think it is the ‘rather than’ which is most telling here. Many people have nailed their colours to the mast to indicate who they side with and ultimately who they will sent code upstream to.
(I don’t know enough about the way patches get merged with main trunk code to know how much cooperation there is likely to be)
My only comment is that Open Office started life as Star Office created in Germany by Star Division.
Open Office is a brilliant product and should not be messed about with outside of the official Open Office team.
Nice overview Rob.
I believe there is something different between LO and those other forks though. All others intent to remain “downstream” receivers of OOo code, as far as I can see. While LO obviously isn’t refunsing to merge improvments either, it does not take their downstream position as granted.
Right, there is no vast differences yet (how could be after 2 months), but LO has cleaned out about 40,000 lines of dead and unused code (which will make merging changes more tricky). It announced to possibly switch to a new Spreadsheet backend (ixion), and LO not OOo will be in Debian/Ubuntu/Fedora/RedHat/Suse, and some *BSDs at least. Ubuntu has put out a job ad for LO development, Novell has about 10 developers on it, Redhat has at least 2. Plus there have been more than 150 *new* contributors to the source code in those 2 months (admittedly mostly doing minor cleanup work). I believe this is a different situation than what most “forks” are in.
Forks are quite technical, what makes an Open Source project great is the surrounding community (not necessarily the coding one). And if you look at how many community supporters switched to LO (while others are still waiting how the whole thing plays out), it becomes clear that this probably not just a short-lived spark.
I wish the OOo all the best, but I believe that LO has more potential. If you looked at e.g. Michael Meeks contributor statistics from 2008, you will see that OOo only had around 35 or so contributors with 4 of them being non-Sun/Oracle employed. Now is that an “Open Source” model?
@speatz, I agree that having an active contributor base of 150 coders would be an amazing accomplishment. Even a solid group of 30 coders, if not all from the same company, would be an impressive project. But I think we should be careful who we count as “contributors to the source code”, especially if we want to make a fair comparison of LO to OOo.
To get to the “150 developers” number it looks like LO is counting those who are providing translation strings, documentation, etc. That’s fine; they deserve acknowledgment as well. But if so, it is then unfair not to count translators, etc., as part of the 2008 OOo statistics. Surely, only 35 OOo developers did not write all the OOo code, all the documentation as well as translate OOo into 39 languages?
Do you know how the numbers look if you did an apples-to-apples comparison, say looking at only people who have contributed C/C++ code patches? I’ve gone through the logs and the LO number shrinks from 150 to approximately 30 coders. That is a respectable number for a new open source project. I think LO should give a good, but realistic number, rather than trying to dazzle with the extremely improbably “150 developers” number.
If I look at who is making the patches. I certainly do see some notable non-Novell code contributors (Thiebaud, McNamara, even yourself), but if I look at the top 10 contributors, 8 of them are Novell/Suse employees. So at the technical contribution level this looks very much like a project that still has concentrated corporate backing, but has replaced Oracle with Novell in the leadership position. Given that, it is not very surprising that Suse/Redhat/Ubuntu are distributing LO instead of OOo. Weren’t they distributing the “Novell Edition” (GoOo) all along? Of course, the sad irony is that Novell has been talking about forking OOo for many years now. But right as they finally gets their shot, Novell is taken over by Attachmate and the substantial Novell/Suse team working on the new LO fork now has an uncertain future. The timing could not be worse.
I wish one of the StarOffice successors had continued Mail and Schedule, and especially their intgration with one another and the rest of the office suite, instead of dropping both from version 6, on. Whatever happened to the code for these modules?
Although Evolution still hasn’t gotten it ‘right’ with a certain kind of recurring appointment that Schedule handled with ease (for how many years by now have Evolution users been waiting for a fix?) the real genius was not that Mail was the greatest E-mail client ever coded, but in the level of integration between the modules.
I haven’t seen any successor, when you could drag a link showing a plain-text name for just about any object inside the suite, on the user’s system, or on the Internet, into a Task or an Event.
And, not to mention having the name of the link displayed in the ‘Also Refer to Frame’ on the right, so the user can enter enter comments about each, in the’Description’ frame, to the left of the listed links. Can’t do that on Evolution, yet, with the same degree of style … I wonder why?
Yes, pre-version 6.0 of StarOffice had problems, but what a pity not all of the modules could have been given the same attention and development as was given to Word, Calc, Impress, Draw and Math.