On Mon, Aug 29, 2011 at 5:57 PM, Dennis E. Hamilton
<dennis.hamil...@acm.org> wrote:
Rob, I completely disagree with embracing everything under ALv2 and exclusive
Apache Way custodianship. I don't consider that a goal. I don't see how that
serves the extended community at all.
I'm not talking about converting existing contributions to ALv2.
Obviously we cannot do that. I'm talking about future contributions.
We'll have a community wiki, yes of course. We have one now. We've
had it since we first started. We'll have a user list, certainly.
But core project functions, such as development, translation,
documentation, user support, etc., should be integrated into the
project, the Apache project. These are core project responsibilities.
Every other Apache project does this. I have not heard a cogent
argument why this should not be true of us.
The thing to let sink in is that this is an Apache project, with an
Apache project community, with an Apache license. We only delay the
inevitable if we continue talking like there will be an independent
entity, using Apache servers and trademarks, but electing licenses
other than ALv2 and using a decision making process and meritocracy
independent of the Apache project and the Apache Way.
We need to start talking about how we integrate and welcome the
community into the project, not just integrate websites. We need to
start explaining the benefits of signing the iCLA and the benefits of
the Apache Way. If we fail to do this, then we fail as a PPMC, and
ultimately as a project.
I accept that is not your position.
As far as I know, the position you express is not a commonly held view, though
whatever view is commonly held is not all that clear.
I'm arguing the logic of my position. You were doing the same, but
now you've veered off to arguing mass opinion.
Furthermore, I don't believe we have the right to appropriate all things at
openoffice.org that are not explicit in the Oracle SGA. It seems that is not
part of the Apache Way.
That is not what I was proposing.
So we need to be careful and respectful of what there is and what has gone
before in taking custodianship of the domain name(s). I think that means
endeavoring to perpetuate those sites in a manner that changes only what we
have to change to bridge to the Apache project. What is not essential to the
Apache project needs to be sustained in its operation for now and we can work
out further arrangements as we go.
OOo also independently raised funds, organized events, gave permission
for trademark use, registered domain names, applied for Google Summer
of Code, collected money from conference sponsors, hosted commercial
extensions on their websites, put Oracle product advertisements in
their install programs, etc. We're not going to continue all of these
things. We cannot continue all of these things. We're an Apache
project. We need to start thinking like one.
There is absolutely no obligation to continue something OOo did merely
because OOo did it.
Apache is not the Internet Archive. We have no duty to curate and
preserve things just because they existed at OOo. As you know,
LibreOffice did not just copy everything as-is from OOo. Why should
we not have similar freedom of action?
Apache is not SourceForge. Unlike OOo it does not work for us to
create nearly 200 independent "projects" beneath our podling that
maintain their own websites, their own wikis, their own code
repositories and their own releases. If there is a project that we
neglect to migrate, because we don't appreciate its value, then there
is nothing that prevents volunteers who worked on that project from
applying to become their own incubation project, or from setting up
their own independent project.
The default action is that nothing happens unless sufficient
volunteers step forward to migrate a service and then integrate it
into the Apache project. This means integrating it not only at the
Infrastructure level, but also as a component of an Apache project,
under the Apache Way and with the reality that a project exists as an
entity within a larger non-profit corporation, etc. That is the
default.
The only alternative I can imagine is if some other party steps up to operate
openoffice.org and Apache provides them the use of the domain name under some
suitable arrangement. That would also be a decent Plan B if we fail to
graduate from incubation.
There are certainly some who are hoping for this. But the hard part
of having a successful alternative service is not the trademark or the
OpenOffice.org domain. The hard part is providing the technical
infrastructure, expertise and volunteers.
Look at http://www.oooforum.org They get tons of traffic though they
are independent of the OOo and get no advantage from the URL or any
official relationship to the project.
Look also at LibreOffice for a similar case. They've made significant
progress, without any use of OOo trademarks, logos or domain names.
Look also at ODFAuthors for another example.
If someone wanted to host an OOo wiki, or community website, or
documentation site, or extensions site, or whatever, they can JFDI.
Lack of the openoffice.org domain name did not prevented the above
projects from setting up successful websites. Why should it prevent
someone today?
In any case, if someone really thinks they need to use the trademark,
logo or domain name then there is always the ability to submit a
proposal to the PPMC. They could also make their own incubation
proposal to Apache. Since Apache, not this project, owns the
trademark, logo and domain name, Apache could then mediate access to
these assets among the Apache projects that request them. That is
also an option we might have at graduation, to move in to multiple new
projects.
I agree that using a private SVN for the non-developer community stuff is not
ideal and would not be permissible if it was all Apache. I'm not wedded to
that. What I found valuable is that it doesn't create any undesirable
contamination of what will be clean Apache project SVN with material that is
not so pure.
Another way to look at it is a web-site/-service version of an Apache Extra,
visible to the world on a non-Apache domain name.
- Dennis
-----Original Message-----
From: Rob Weir [mailto:robw...@apache.org]
Sent: Monday, August 29, 2011 14:33
To: ooo-dev@incubator.apache.org
Subject: Re: [DISCUSS] RE: SVN and bringing the total end-to-end OOo project
under Configuration Management
On Mon, Aug 29, 2011 at 5:09 PM, Dennis E. Hamilton
<dennis.hamil...@acm.org> wrote:
I've been mulling this over and I am wondering about another way to look at the
problem, building on Eike's suggestion too.
This is not a proposal. It is too high-level and not concrete enough with a
viable roadmap. We need to see if we can find a consensus in principle and
then see what kind of roadmap would have http://openoffice.org continue in
operation. The goal is as little disruption as necessary to achieve rehosting
and sustained operation on behalf of the extended community and also create an
effective firewall between the Apache project and non-Apache community efforts
such as the NLC activities.
"As little disruption as necessary" and "sustained operation" are
opposing desires. I think we should seeking to optimize the project
as an Apache project, not merely transport everything as-is.
Another point is that, regardless of the legacy of the project, Apache
2.0 license is not antithetical to community contributions. We should
be requiring it everywhere, for all contributions. If we wanted to be
true to the past, then we'd require copyright assignment to Oracle,
right? Obviously, moving from that to ALv2 is a big step forward. So
the goal should be to integrate the community into the Apache project,
not try to create firewalls to prevent effective collaboration. A
common license is the best way to encourage effective collaboration.
- Dennis
BASIC DIRECTION
I think the community material should not be underneath any of the existing
svn.apache.org/repos/asf/incubator/ooo subtrees (not site, not trunk, etc.).
My suggestion is that we use one or more new subtrees. An easy way would be to
have a single subtree ooo/community with site, wiki, and forums under it. Or
they could be higher-level.
I also think we should consider placing one or more of those on a private SVN
tree. The private repo would only visible to committers at this level
(although it probably means that commit messages go to ooo-private rather than
ooo-commits). WE ALREADY HAVE A PRIVATE SVN repo that could be expanded for
this. [I'm not sure this makes sense for backups but there may be other
artifacts that would belong under SVN for the forums and wiki.]
This goes against transparency. The only things that should be in our
private SVN are confidential things, such as lists of committer email
addresses, etc.
CONSIDERATIONS
The reason for private SVN is to avoid public disclosure of community
openoffice.org material via anything other than the web site, the forums, and
the wiki themselves. It defers having to deal with current content that is not
sanitary with regard to Apache licensing, while continuing to have that
material available to the extended community under the conditions of its
original contribution. It also assigns custodial responsibility to authorized
project committers, resolving a PPMC duty to the ASF.
The only firewall we need are diligent committers. Remember, there
are many things out there in the wild, wild internet that we must not
cavalierly check into the SVN source tree. We rely on committers to
be careful everywhere and always.
And we don't want the community wiki,etc., to always be verboten to
project use. We want, over time, for there to be an increasing amount
of ALv2 material that could be used in other contexts. I think we do
a disservice if we merely set up a situation where the community can
continue to contribute material that is in a contaminated pool of
variously and ambiguously licensed content. We owe it to the
community to bring some clarity to this problem.
Note that this means we should pursue the proposal of Kay Schenk (and, I
believe, Jean Hollis Weber earlier) to *then* migrate the community web content
to the community wiki. One important benefit is removal of the need for Apache
committers to maintain community material. It is difficult for the community
to submit issues and patches against the private bits, so we want to make those
bits as few as possible. This is also a way to retain the NLC sites. If there
is any special custodial relationship needed there, we can work that out
without complicating the Apache development project and sites on the apache.org
domain. In particular, NLC material and public Apache project materials would
never be comingled.
Finally, private or not, the separate tree for the web parts is now amenable
for serving up as a different web site under the openoffice.org domain. Having
the infrastructure and the repo be separated (better: private) avoids confusion
with Apache-licensed content and project releases as well.
When the migration is completed, I suspect that the OOOUSER Wiki could be
merged over to the OpenOffice.org wiki (or not, since it seems to provide an
important development-side focus). http://incubator.apache.org/openofficeorg/
and the eventual http://openofficeorg.apache.org sites (though I prefer
http://ooo.apache.org) would maintain the Apache project development face and
http://openoffice.org would provide the community face, with redirection to
developer-focused apache.org locations (such as for issue tracking, etc.).
-----Original Message-----
From: Dave Fisher [mailto:dave2w...@comcast.net]
Sent: Monday, August 29, 2011 09:02
To: ooo-dev@incubator.apache.org
Subject: Re: SVN and bringing the total end-to-end OOo project under
Configuration Management
On Aug 29, 2011, at 8:30 AM, Michael Stahl wrote:
On 29.08.2011 16:40, Terry Ellison wrote:
Thanks for comments. Rob. I was hoping to get your and others thoughts
on this TLD structure issue: Where do we plug the wiki, the forums, the
other website services into our svn hierarchy (where XXXX=the relevant
service):
* incubator/ooo/site/trunk/XXXX
or
* incubator/ooo/site/XXXX/trunk
or where?
There's no clear slot in our current TLD structure. I've put my
responses on the rest below.
i don't understand at all why "site" contains "trunk", does anybody
really want to branch it?
In preparation for a release!
Regards,
Dave
On 29/08/11 14:59, Rob Weir wrote:
2) Our customizations occur in a branch
Tried this before on projects. It really doesn't work. There are
~2,500 files of which we update about 20-30 with a single patch file.
If we do it the way you suggest, we would have a huge bulk of revisions
every phpBB release. It's a lot easier to keep the build script and the
patch file under CM and then we only have two files to update every
release.
perhaps a patch tracking tool like "quilt" would be appropriate?
http://savannah.nongnu.org/projects/quilt/
it allows to have not just a single big patch but multiple patches, each
one containing one "logical change".
then the patches and quilt metadata can be put into SVN.
(i have been using a versioned HG Mercurial Queue against the OOo repo,
which is quite similar in approach...)
regards,
michael
