Re: Need Leadership

["Jason D. Clinton" <me jasonclinton com>]

On Sat, Jun 21, 2008 at 1:26 PM, Havoc Pennington <hp pobox com> wrote:

2008/6/21 Jason D. Clinton <me jasonclinton com>:

> In my opinion, whatever "The Next-Gen Gnome" is, it isn't going to happen
> until we really, really have a deep maintenance cycle going on here. That
> means fixing a Handful of Giant Warts on our maintenance process:

I bet the next-gen gnome will happen when someone writes it. I would
suggest people think in terms of getting something going by
themselves, and once it's at least roughly usable, think about
recruiting 2 or 3 or 5 other people to the project. But getting
hundreds of people to agree up front isn't too likely. Think 5 not
500.

I'd suggest also that "next-gen gnome" is a bad framing. It's the same
broken mindset as "GNOME 3.0." GNOME _the desktop_ does not need a 3.0
or a next-gen in particular. I think most of the current users, and
current involved OS vendors would basically be against major change in
the target audience of today's GNOME desktop (current Linux users). I
mean, even if individuals at those companies are in favor in
principle, it's not their day job and their day job has major
pressures to focus on the current audience.

I apologize for I intended this thread to be about getting our ducks-in-a-row with regard to 2.x maintenance. It's my own fault for piggy-backing off of Decadence Thread of which so much of that has been about Gnome++.

I whole-heartedly agree with you that setting some lofty goal that everyone has to buy-in to is antithetical to a healthy development process. But, again, I got us off on the wrong foot by including the above sentence in my post. Hopefully this thread can be more about doing the things that make incremental improvements even easier and radical new experimentation trivial.

Allow me this one concrete example: in Bugzilla, there's a bug open against Iagno in which a non-native English speaker has re-written the entire game from the ground-up adding multi-player; the diff is 4,000 lines. I really like where it's going but unfortunately, as someone new to FOSS he really didn't understand the "contributing patch-sets" custom that we've come to all agree on. Now, if we were using DVCS, I could have this new version of Iagno incorporated in a mater of days because all the changes would be incrementally documented with rationale. Instead, we're looking at months of work to break his changes in to peer-reviewed patch-sets that are documented and transparent from the perspective of the ChangeLogs. (ie. "blame" outputting a person's name and the rational for a change on a per-line basis).
 

Don't misunderstand my point here: I don't think anyone should "cave"
to current users or commercial pressures. I do think that it's
impractical to ignore almost everyone currently working on or using
the software, though. You just can't fight that momentum. It isn't
even correct to fight it. There are lots of users there with
expectations.

The goal is not to randomly churn up the GNOME desktop as it exists
today (window manager, session manager, panel, etc.) - that thing
should just keep evolving in incremental fashion, getting better all
the time for the people who use it.

+1
 

The goal should be to find all the new directions, and see if GNOME
can start to be about those too. Right now there are lots of new
directions in mobile, set-top boxes, EeePC-type thingies, for example.
Why does the front page of gnome.org still say "GNOME offers a
desktop" - excluding these directions from "GNOME"? That's a problem.
The "GNOME stack" potentially has much wider applicability.

Excellent point.

Is there a list somewhere of the rich ecosystem of consulting
companies and libraries and products that build on the GTK/GNOME
technology stack? Why isn't gnome.org positioned as about that list,
with "the desktop" as only one of the things on the list? I bet a
solid fraction of our community isn't even primarily focused on "the
desktop" anymore... could be a majority even. But the
Fedora/Ubuntu/Suse framing of GNOME-as-desktop remains dominant
despite all the smaller companies who are doing other stuff.

Yes. And your concerns are right. For example, Cristian Persch has been working in my module on making games run on Maemo. Others have worked on Windows ports. I, personally, love this work. What I would love to see, however, is the development process and release schedule take in to account the possibility that modules in Gnome are not all 100% aimed at the desktop. For example, there hasn't been enough emphasis on release goals that have targeted making our modules more portable. It seems we're stuck thinking:

  "What will this look like in the Release Notes when Gnome 2.x is release?"

As opposed to what would perhaps be a more healthy:

  "How does this make my module more universal?"

Setting release goals and rallying the troops to them is very much a leadership issue.

GNOME could be instead of "a desktop" something like "a development
community for open source user experience technologies, on all
platforms including Linux, Windows, and mobile devices" or something
like that. I don't know. Some broader goal. Describe the community,
not one of its products.

Precisely. 

 

I don't think git or mono have much to do with anything. If you start
a 1-person or 5-person project, you can use whichever vcs or language
you want. And the bottom line is that anyone who blames their version
control system for inability to get stuff done is not the kind of
person who gets stuff done. git did not exist when the vast majority
of the tens of millions of lines of open source code were first
written. (Maybe GNOME or parts of GNOME should switch to git or mono,
not debating that, I just think it's bogus to say either one is truly
what's blocking innovative new projects.)

People should also think creatively about how to get things done,
beyond just writing code. Miguel and Nat back in the day created
International GNOME Support, and that was only one of the companies
involved. The GNOME Foundation was created as well. Today, you see
plenty of small companies around GNOME. These are all people thinking
not only about what code to write, but how to add more time and more
people focused on the project.

I agree. The Mono suggestion is in that vein but also brings us back together from what appears to be a dangerous and unnecessary rift between Novell and everyone else. I hope that we can put the past behind us to the extent that their work has no known patent issues. (And I know that the vast majority of it does not.)

No one should think of the 3 items that I've listed to start this thread as panaceas. They are merely actionable items that aren't getting enough attention that would make *new* developers coming to Gnome happier. And that's why it matters.