Re: [tdf-discuss] Re: [Libreoffice] Proposal to join Apache OpenOffice

On Sun, Jun 5, 2011 at 6:00 PM, Jim Jagielski <jim@jagunet.com> wrote:
> On Jun 5, 2011, at 10:15 AM, Simos Xenitellis wrote:
>
> > What can the Apache Foundation provide to OpenOffice?
> > 1. You start with zero community and you alienate the LibreOffice community.
>
> I would submit that "zero" community is somewhat of an understatement.

You start with a zero "OpenOffice" community. The Apache Foundation had
no prior involvement in the development of OpenOffice.org due to the license
not being permissive.

> As far as alienation, if the LOo community feels alienated, and I'm
> not saying that they don't have every right to, why is it directed to the
> ASF which never sought this donation, and from the get-go has tried to pull
> in the LOo organization?

It's actually LibreOffice (LO).
Whether the ASF tried or not to pull in the TDF, or how well it tried,
is something that I do not know.
Apparently, these are private discussions between the two organisations.

From what has been alluded to on the other mailing lists, apparently
Oracle wanted to transfer
the full copyrights of OpenOffice.org to the ASF. For a donation that
the ASF never sought,
it would have been ideal to accept the donation and then transfer the
full copyrights to the TDF.

> > 2. You will start building a community at some point in the future in
> > some unknown way.
>
> Please read the various posts and sites regarding the Apache Incubator
> which describes how this is done, and has been done, quite successfully,
> for loads of projects.

You are talking about how to build a smallish group of developers.
This office suite requires a bigger community, including support network,
localisation, usability and testing teams.

> > 3. You are developers and can currently only deal with developer needs.
>
> We are users and developers. Anyone with even a rudimentary awareness
> of the ASF and how ASF projects work should realize that. In fact,
> the very 1st ASF project, the httpd server, should clearly indicate that
> it was as *users* that we used our developer skills to keep the project
> going.

The most well-known Apache project has a single user support mailing
that dropped to about 350 mails per month.
The ASF does not do software like office suites that end-users will
use on a daily basis.
Perhaps a library or non-GUI elements of an office suite, but not the GUI.

> > 4. Your infrastructure is based on Subversion (SVN) which will make it
> > difficult
> > for other to share code. Git is not even in the immediate plans.
>
> git is currently being investigated. svn allows others to share code.

It was said in the incubator list that it would take several months before
a DVCS is selected, and apparently it's open whether it will be git or
something else.
LibreOffice did the hard work converting into Git and cleaning up.
So, the ASF plans to do the whole work again from scratch, not now
but in several months in the future.
How can you merge patches between the two repositories if you have to deal
with minor changes due to the different cleanup (will ASF cleanup the OOo code)?

An office suite is of great interest to governments around the world.
They would
want to start their clones of the code and contribute. How would they deal with
the ASF version of the code?

What about ASF developer training to start using a DVCS?
Those IBM Lotus Symphony developers, do they use internally a DVCS?
Do they have experience working with the community? Because several Sun/Oracle
engineers did not have such training.

> > 5. You are happy to get going with 20-30 core developers.
>
> And why not?

Because an office suite is of great interest to governments and organisations
around the world. You would want to attract developers from around the world.
The IBM Lotus Symphony developers apparently are around or over 20 (is it?).
How is developer diversity going to work?

> > 6. The Apache Foundation hosts over 150 projects and I fail to see
> > any important user-centric software like OpenOffice.
>
> Agreed, if by "user" you mean "desktop end user." And the ASF has
> been quite upfront in saying that this is an area where TDF has
> some clear areas to provide insight/help/guidance, etc...
>
> > The essential need for the Apache Foundation involvement in this appears to be
> > so that IBM can continue to offer a proprietary product, IBM Lotus Symphony,
> > License Agreement at http://pastebin.com/uqbUTRg5
>
> No, the essential need is that Oracle wanted someplace with a
> proven track record to donate the code to so they could then
> be rid of it. The essential need to the community is an open,
> well-established entity that is (hopefully) able to help the
> entire community to cooperate and collaborate on such an important
> piece of FOSS code as OOo.

The whole community is behind the TDF and LibreOffice.
Apache OpenOffice is starting from scratch, and has to do lots
of work to catch up, work that has already been done in LibreOffice.
The process to empower new developers has started, and there are
numerous people that did their first commit. This is growth potential.

If I were to speculate, the choice for the ASF would be that
1. it's permissive
2. it's IBM dominated

> > Is IBM is trying to replicate what Sun/Oracle had with StarOffice,
> > putting just enough resources
> > for their own needs in order to ship their product?
>
> One could ask the same of Novell, but in any case that is
> immaterial to the point. By building a healthy community around
> the project, what IBM/Novell/Foobar "tries" to do is moot.

LibreOffice is copyleft, so it's only IBM in the picture.

If IBM has no interest in a proprietary, closed-source office suite,
then there is nothing that keeps them from joining the TDF and LibreOffice.
They would get a fully localised package in many languages,
support groups that help users in using the software.
Git repositories are ready, user groups around the world are waiting
to help end-users.

> > The Linux kernel is an amazing piece of software that it used in 92%
> > of Top500 supercomputers,
> > all sort of servers, mobile phones, most TVs and routers.
> > And still, there is a single Linux kernel project thanks to the
> > copyleft license.
> > Everyone works on Linux because they cannot keep away their own contributions;
> > they have to share them with the community.
> > Even the ARM architecture, where each ARM licensee went their own way,
> > is going to get its cleanup.
> > Because the code for all of them is already in the Linux kernel repository.
>
> One could point to the success of AL codebases in the same way.
> Ideological stances on licenses have a tendency to get in the way.

The issue of not fragmenting the developer community is a pragmatic one.
With a permissive license, and a huge interest around the world for an
office suite,
it is more likely than not that fragmentation will take place.
IBM Lotus Symphony is one such example.

> > IBM makes money out of Linux by providing services. And IBM is even a
> > top contributor to the Linux kernel.
> > Would IBM hypothetically prefer to have the Linux kernel developed
> > under the Apache Foundation?
>
> immaterial.
>
> > OO/LO are in this critical point where they can repeat the Linux
> > copyleft success story
> > and help ODF dominate the document formats.
>
> Even FSF admits that when there are competing "standards", a AL license
> is the best choice, even compared to a copyleft one.

Let's read the document you cite,
http://www.gnu.org/licenses/license-recommendations.html

A permissive license is recommended/suggested in two cases, when
a. «very small projects»
b. «projects that implement free standards that are competing against
proprietary standards,
such as Ogg Vorbis (which competes against MP3 audio) and WebM (which
competes against MPEG-4 video)»

I cannot fit OpenOffice in any of these criteria.

> > OO/LO is a complicated piece of code and will probably require big
> > architectural changes.
> > Having an Apache OpenOffice and a LibreOffice will slow down progress
> > in major changes.
>
>
> One could also say that having both cooperate would greatly speed up
> progress in major changes.

Can you imagine cooperation? The ASF did not want OpenOffice in the first place,
the ASF cannot accept copyleft code even in these extreme circumstances,
the ASF will take time to get a DVCS (hopefully git) and retrain developers,
the ASF will start from the old OpenOffice.org sources not being able to use
the cleaned-up LibreOffice sources.

Why didn't Oracle dump OpenOffice to the FSF instead?

Simos

-- 
Unsubscribe instructions: E-mail to discuss+help@documentfoundation.org
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://listarchives.documentfoundation.org/www/discuss/
All messages sent to this list will be publicly archived and cannot be deleted