incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Simon Phipps <>
Subject Proposal for OpenOffice Incubator strategy
Date Fri, 03 Jun 2011 00:12:40 GMT
TL;DR version: I think I see people talking past each other for a bunch of reasons, and I have
a compromise proposal that might make things easier. It's at the bottom, and explained in
some detail in the middle.

Before I start I will introduce myself. I was at Sun for a decade, more than half of it working
on the open source portfolio including Because of that I've a deep awareness
of the history of and a good idea where most of the bodies are buried. I count
many members of Apache and of the OpenOffice community as friends, both on the
Project (where I'd apply for membership if such a thing existed) and the LibreOffice Project
(where I applied and was accepted as a member). Today I am not part of any leadership for
any project that's relevant here and I speak only for myself (and especially not for OSI where
I am also a director). My dear wish for the last decade has been to see OpenOffice become
an open and meritocratic open source project, and that goal has slipped tantalisingly from
reach more than once. 

I've tried to keep up with all the developments in this saga, as well as talk with all the
folk I know how to reach informally to understand its dynamics. My apologies in advance if
I've missed important details.

There's a gap
It seems to me that there's a big gap in the conversation. There's lots of talk of the vision
for the future, of how things can be once a critical mass of developers reach Apache, of all
the effort IBM wishes to invest in a resurrection. But it overlooks the fact that OpenOffice/LibreOffice
isn't dead yet. There are tens of millions of people all over the planet who are using either (especially on Windows) or LibreOffice (mainly on Linux) and who rely on the
combined work of the (now divided) existing community to deliver a steady stream of new releases
incorporating bug fixes and improvements. There are also many people depending on downstream
versions of these two projects. OpenOffice/LibreOffice is an existing, running machine with
an enormous, incredibly important ecosystem. The big gap represents our collective responsibility
to these people in the interim. This isn't just a neutral tarball we have to consider.

Just as the Mozilla project was reborn by the creation of Firefox, so I have long been a proponent
of an OpenOffice project that similarly does a new, vibrant thing. I truly hope that Apache
can be the home to a project that does this great work that groups of us have tried and failed
before to get started - groups that include most of the new names Apache is seeing show up

All the same, the world wants OpenOffice to carry on while all this revolution is happening.
There needs to be a build system, a distribution system, a contributor methodology and more.
All this is needed not for some future, re-invented project but right now. Millions of people
globally know what is, and they want it to keep going.

It seems to me that all the talk here mixes two different sets of requirements, both of which
I support. There's no doubt that needs a Firefox-style rebirth. There's also
no doubt that the distribution generally known as and also widely known as
LibreOffice is still crucially important.

I suggest we discuss how to do both. We need to both keep the wheels turning so that Windows,
Mac, Linux and other platforms get regular releases of updated code. We also need to devise
a way for the New Thing to happen. So I suggest we explore something like this:

1.  Apache should accept the contribution of both a copyright license to the code from Apache
(and I suggest checking it's the full, multi-branched source including all the in-progress
contributions that are in a CWS and not just the release) and the trademarks so that those
resources are secured for the community.

2.  This incubator project, which sets out to be the "Firefox of OpenOffice", should proceed
pretty much as described, but under a name other than OpenOffice (just as Firefox got a different
name). Something like "Apache ODF Suite" that describes the intent to be the core code of
a fresh start. Picking an alternative name will help avoid those millions of current users
getting confused, and I suspect will cool down some of the emotions in this discussion. I'm
sure Rob and the others behind the proposal will be able to populate a podling to get this

3. Given that a substantial part of the effort that the LibreOffice project has committed
has been the creation of an open repository and build system coupled with an effective international
distribution system, I suggest that we collectively ask LibreOffice to take on the task of
"business-as-usual" for OpenOffice, so that the Incubator project can focus on rebirth and
not get swamped in the minutiae of "business as usual".

This new OpenOffice/LibreOffice project will also need to be the upstream for many of the
non-public projects listed in the proposal (as well as being a downstream for the code from
the incubator project when it graduates eventually). That will be fine for some of them, and
others will probably need to consider engaging at the new "business-as-usual" project to make
things OK. I suggest Apache ask this project to describe itself as "OpenOffice" so that everyone
knows both what it is and where to go for the familiar code they are expecting.

I realise there are some big social challenges in this proposal and it will take acts of grace
from across the divided community, but given the alternative of either trying to invent a
completely new set of infrastructure in the Incubator to sustain business-as-usual or using
what already works it seems worth asking for conciliatory grace from the members of the two
sides of the existing public community. 

This is purely my own thoughts, and there's no doubt room for improvement although I have
run it past a few wise friends before posting it. But I suggest that without this clear demarcation
of "new-project" and "business-as-usual-project" it will be very hard to disentangle the two
sets of needs and fulfil the worthy objective at the start of the proposal, "Both Oracle and
ASF agree that the development community, previously fragmented, would re-unite
under ASF to ensure a stable and long term future for". 

Simon Phipps,

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message