Digressions on AOOo
I have one quarter of Roman blood, and I have spent my youth in Rome where I have attended the primary school. I can still speak with a decent Roman accent, so AOOo – the short name for Apache OpenOffice – has the same sound of the shout used in Rome to grab attention: “ahooo” (where the H has no sound at all).
AOOo, at the moment, has not produced anything worth of an “ahooo”. On the contrary, has raised a number of concerns in the media immediately after the announcement. Concerns shared by an increasing number of members of the OOo community.
I sympathize with IBM, as since the inception of AOOo Big Blue has been object of many of the blames that should have been directed to Oracle.
We should never forget that Oracle has generated most of the problems that have frustrated the community and led to the birth of The Document Foundation, then has shut down the entire OOo project with a single press release and subsequently fired around 120 people, and has eventually decided to choose Apache Foundation over The Document Foundation as the new house of OOo (generating a second wave of nasty feelings inside the community).
IBM, as the main corporate sponsor of AOOo, is in a quite unconfortable situation (because many have perceived that the company has been instrumental for the move to Apache Foundation) but should not be blamed for Oracle faults.
Rob Weir is working hard to keep together what is left of the OOo community and push forward AOOo. I believe it is important for IBM to engage with the community and contribute code, as a failure of AOOo would go against the interests of open standards, software freedom, choice and – ultimately – The Document Foundation.
Looking at the list of committers, it is quite clear that the number of “real” developers is quite small (in relation to the total number of committers) and the number of “real” developers who have a decent understanding of OOo code is even smaller (and probably decreasing over time, if some of the former Oracle employees will be finding a job outside the OOo ecosystem).
In addition, the majority of the committers who know the OOo code (former Hamburg developers, plus IBM and RedOffice chinese developers) have not been very active so far in the discussions.
Today, IBM has the opportunity of hiring several of the former Oracle employees – who are looking for a new job – to work full time on AOOo, in order to speed up the process (before releasing the first AOOo version of OOo, based on OOo 3.4 Beta, the project needs to get rid of the many bugs affecting the code, and to replace the components licensed under GPL or LGPL: quite a significant amount of work).
IANAD (I am not a developer), and therefore I cannot express a professional opinion, but I feel that IBM should hire around a dozen former Hamburg developers, in order to have a number of sponsored hackers similar to The Document Foundation. Otherwise, AOOo will risk to lag in a significant way behind LibreOffice or even fail, and no one will be happy and ever shout “ahooo”.
Italo,
Thank you for not blaming IBM. Let’s agree to move forward.
My personal view is that there is in truth, one OpenOffice community. It is also true that there are within this community different projects with different names and so forth. LibreOffice, Symphony, EuroOffice, etc. Ultimately, it will make sense for the project members to converge on one development effort, without negating any particular project’s goals for differentiation. This is what IBM has announced and will do. We hope others will equally share in the effort and the benefits.
Apache OpenOffice has the potential to support all. The Apache Software Foundation is about volunteerism, no less or more in the same spirit as TDF. Effort invested in Apache OpenOffice will benefit LibreOffice.
Give it a chance.
Hi Don, I don’t see why I should have blamed IBM.
Others have done it, and I think they were wrong or inaccurate or superficial.
On the other hand, I still can’t understand how AOOo will survive without a large development community (believe me, it is difficult to attract, motivate and grow a community of developers).
I know that Apache Foundation is about volunteer work, but – in my paramount ignorance as a developer – I don’t see how a small number of volunteer developers can be able to manage such a large “beast” as OOo.
I hope I will be disproved by the course of events.
Why can’t Apache OOo give the project back to TDF?
Why can’t Apache OOo give the project back to TDF?
Because they shouldn’t, at least not at this point in time. Why? Well, unlike TDF, Apache is actually able to resolve the licensing issue – and they likely will do that because they plan to put AOOo under the Apache License, which is much freer than LGPL (no more copyleft!).
TDF itself promotes making the licensing freer (IIRC, they favor triple-licensing GPL/LGPL/MPL, with at least MPL being freer than LGPL), but in effect they only can do that for NEW code contributions because they cannot change the LGPL-only licensing of Oracle’s codebase.
However, due to the copyright donation from Oracle, Apache IS able to change the LGPL-only licensing once they stripped all “foreign” GPL/LGPL code dependencies. (Which is also some tremendous amount of work, but that is another story.)
So in effect, I’d really like to see Apache release the entire codebase under the terms of the Apache license and remove the harder LGPL requirements. Once that is accomplished, I am pretty sure that numerous companies will join in and support the further development (it is my exprience that many commercial comanies are not willing to support LGPL projects, while being much more comfortable with the Apache license).
And finally, this will benefit the entire OOo community including LibreOffice, because (to my knowledge) the Apache License permits being used in a GPL/LGPL/MPL-project. So LibreOffice could finally put their own favorite licensing onto the entire project, just because of Apache’s effort.
I don’t think that companies will jump on AOOo once the code has been released under the Apache License, as they didn’t jump on OOo when the code was released under the permissive SISSL (only IBM did). On the contrary, the number of companies supporting LibreOffice is growing steadily, and it doesn’t look that the LGPL is a problem for any of them.
Don – when you say:
> Effort invested in Apache OpenOffice will benefit LibreOffice.
Why do you think that ? Given the really rather substantial divergence of the code-bases, I (personally) don’t see a wholesale / regular re-basing of LibreOffice on Apache OpenOffice as realistic. Some cherry-picking, perhaps – but… the LibreOffice codebase has (by now) some pretty vast cleanups applied that improve things a lot but make merging harder. Indeed, beyond the licensing improvement, it is hard to think of a benefit that Apache OpenOffice brings to LibreOffice.
@italovignoli:
they didn’t jump on OOo when the code was released under the permissive SISSL
I admit I didn’t know OOo was ever released under a permissive license, and I don’t know the exact terms of SISSL. When was that, and why was the permissive licensing withdrawn in favor of LGPL?
the number of companies supporting LibreOffice is growing steadily, and it doesn’t look that the LGPL is a problem for any of them
Not for them, sure. But what about all those companies that do *not* support LO? My experience is that BSD-style licenses (or stuff like Apache/MPL/CDDL) are by far more attractive to companies than GPL/LGPL. But of course, you may have different experiences.
@Michael:
beyond the licensing improvement, it is hard to think of a benefit that Apache OpenOffice brings to LibreOffice
You’re probably right. But still, the licensing improvement alone would be such a substantial benefit that it is worth the whole thing.
OOo 1.0 and 1.1 were released under SISSL, which was dropped when it become clear that the permissive license was not bringing any advantage to the software. Permissive licenses might be attractive for server or infrastructure software, when you usually do not have a widely available free version. For instance, Apache Server is not accessible – technically – to a “normal” end user, so a proprietary version packaged in a specific way might be interesting for many corporate end users who do want to avoid the hassles of developing a version optimized for their needs. With LibreOffice and OOo, though, we are talking about a software which most end users are able to install on their own. This eliminates the need of a proprietary version, and the history has shown that the proprietary StarOffice and Oracle Open Office have never been able to generate any visible turnover (and have not attracted a single company). Not even Lotus Symphony, which is proprietary but available for free, has been able to achieve a visible market share in comparison with OOo. Unfortunately, Sun has never understood that it would have been better to develop and sell value added services around a free software than to sell licenses of a proprietary software. OOo has attracted many companies – especially small ones – because they were able to create – albeit in a totally unstructured way (because Sun never invested in the creation of a ecosystem) – a value added business around it. End users, including corporate end users, were migrating to OOo because it was free software, and were buying value added services on top of it (from the small companies). None of these companies – some of them very large, up to 100.000 seats – ever asked for a license of Sun StarOffice. End user software is peculiar, in comparison with server or infrastructure software. In this sense, OOo and LibreOffice can be considered unique case studies, which contradict many of the software industry beliefs. Server or infrastructure software is a different story, but the two cannot be mixed. By the way, LibreOffice new developments are licensed under LGPLv3+ and MPL, and when OOo will be re-licensed under the AL we will be able to change the entire code base to LGPLv3+ and MPL (but I am pretty sure that MPL will be ignored by most users, including corporate ones).
Thanks for all the interesting background info! I agree that proprietary flavors of OOo/LO have never been as successful, but I’d still question if value added services on top of a free application are a promising business model either (experience teaches that “free core” has largely failed).
By the way, LibreOffice new developments are licensed under LGPLv3+ and MPL, and when OOo will be re-licensed under the AL we will be able to change the entire code base to LGPLv3+ and MPL (but I am pretty sure that MPL will be ignored by most users, including corporate ones).
I already wrote that, and in fact that is what I consider the best thing that Apache can (and hopefully will) do to LibreOffice. I respectfully disagree with MPL being ignored, and I’d even recommend triple-licensing LGPL/MPL/AL as soon as Apache makes it possible. So that any user (or potential supporter) can freely decide which OSS license best fits his individual needs.
Let’s see what happens at Apache. And again, thanks for all the background information!