Thursday, February 09, 2006

Where to from here?

I am continually amazed at the number of positive comments I get on my blog articles: it seems that I am not alone in some of my thoughts! That is always reassuring :-)

The things that I have been blogging about, the comments that I have been receiving, and the recent controversy on the desktop development list (sparked by discussion of the recent showcasing of Novell Linux Desktop) have all started me thinking about this:

Has GNOME lost its way?

By "GNOME" I mean the GNOME "community" as well as the bunch of zeroes and ones that are currently chugging through my computer's CPU. I do not mean to imply that things are bad, it just seems to me that we seem to be somewhat aimless and fragmented. I am not suggesting that we need a benevolent dictator like Linus Torvalds or Larry Wall, or that we need more structure or formalism. (It may be that we do, but I remain to be convinced on that score.)

What I am am suggesting is that we need to articulate our shared values and goals a bit more explicitly. (I think the place for this is the eagerly anticipated but oft-delayed new developer.gnome.org site.) In particular, we need a longer term plan that just the next six months (the next realease) and more concrete than the mythical Three Point Zero.

Can we start to think about 2.16, 2.18, 2.20 and 2.22 and publish these plans on live.gnome.org? Can we nail down a few things we want to achieve in the next two years and track our progress toward them? I think that this would help unify us and give us a common goal much more than anything I see in public channels right now.

12 comments:

bkor said...

http://live.gnome.org/RoadMap

There is one 2.16 task there and loads without a target. For Bugzilla I wouldn't be able to say when exactly stuff would be implemented, but I do know what things I want to do / change / improve. And Bugzilla is not even part of GNOME (Bug-Buddy is separate).

late said...

As honourable as your intentions are, I'm not if your goal of a unified plan for the future is really possible. The Gnome community really is fragmented, much more so than the examples of Perl or Linux that you cited. Novell, Red Hat, Canonical, and Sun all plan based on the desktops they want to deliver to their customers -- and each of these goals will be different according to the different needs and opinions of these groups. There is no central organizing figure to plan what Gnome will be like Larry or Linus can for their area.

But this statement, while uncomfortable perhaps, is not utter disaster. What is important is that these organizations and the non-affiliated individuals working on Gnome can identify areas of common interest and that they work together when their goals coincide. It is not really aimless, it's just that the fragmented nature of the differing goals means that there are several different aims as being part of what constitutes Gnome.

It is very natural that there will be some friction when the goals are different, and there will be heated opinions tossed around and sometimes even code duplication. But the challenge is for the community to realize their common interests (ISV support, performance, usabality, etc) and try to work past the various areas where they diverge (languages, release timing, theming).

Ploum said...

It juste gave me an idea.

I suggest to explicitelly tell that some "project" can choose to skip one release to focus on a more long term goal.

For example, let say that Abiword want to implement some big changes. They can say :
"We will make big big changes. The stable release for 2.16 will be the same that the one for 2.14, a few bug fixes excepted. We want to focus on the 2.18 release to make some major breakthrough"

It's already possible, I don't know if it is useful. But telling this in a more "official" way can maybe help some developpers...

Anonymous said...

There is a roadmap already:
http://live.gnome.org/RoadMap

There are more productive ways of asking than saying "Is the sky falling?" No, it's not.

Murray

Anonymous said...

Gnome offers many different applications which is fine, but is not hard to find some discord between the developers, for example:
Muine, Rhythmbox, Banshee, Quod Libet....(I can bet that there're more) why they don't agree on something?, why they don't just join their efforts to do a good program? (as KDE has its Amarok, a really good program), instead of doing almost the same thing over and over again they should decrease their ego and give their work to the most advanced project or colaborate with their new ideas and codes to it and not "create" a program that does almost the same to the previous one.
Thanks

Tom von Schwerdtner said...

http://live.gnome.org/RoadMap is, IMO, pretty lousy.

1) There is no way in hell that all significant changes are listed there. An easy example is the removal of icons from the background settings UI. Maybe people wouldn't get so pissed about these changes if there was any indication to the community that they were coming and why.

2) There is very, very little information about the planning. Half of the entries are one-liners with a listed responsible person and no link to additional information. Where is, for example, an ISV supposed to go to find how how their software or future plans will be affected? What modules will be affected and how?

Basically, if you want to know what is going on with future development you have to subscribe to all the -dev mailing lists with software in the release, monitor all the relevant IRC channels, follow all the important bugs that come into bugzilla, go to all the conferences, hang out with all the main developers and keep up-to-date wit planet.gnome.org and follow all the comments on the relevant blogs.

At least it feels that way.

It seems to be a semi-annual occurrence that some apparently hidden discussion leads to a decision that pisses a lot of people off. The justification for this seems to be that someone knows better, and maybe they do (I think they usually do), but the fact that these changes seem to come as a suprise to many is IMO a big problem.

My $0.02 USD anyways...

bkor said...

tom: removal of icon of the background settings UI is a change made by one developer. It is not a general idea. It shouldn't belong on a RoadMap.

That more stuff could belong on the RoadMap, I'll agree.

It seems to be a semi-annual occurrence that some apparently hidden discussion leads to a decision that pisses a lot of people off.

And so a discussions starts and it is resolved eventually. Having a discussion with everyone who might have an opinion makes no sense. You do that with a select (but relevant) group of people. We should be able to trust the maintainer(s) to do the right thing. Sometimes a change will be for the bad, so be it. If a maintainer doesn't fit within Gnome, the solution shouldn't be to add more rules / bureaucracy.

Tom von Schwerdtner said...

bkor: Dammit, you're right. I started off with what I think was a valid point and I slipped into this "dammit, why wasn't I consulted" line of thought.

I do think my basic point was valid though, that the roadmap is incomplete and lacking in information.

What I would love to see is something like this (plone.org) for GNOME. Does it seem outlandish or excessive?

bkor said...

Well.. I'd want lgo/RoadMap to be for the big changes (there should be more on them). Showing every bugreport might be too much.

I did had something in mind like the plone.org page (extension of browse.cgi), but that would be product-specific. For a Bugzilla-wide listing the version and target milestone of each product would need to be linked to a GNOME version. There is no connection currently, and I can't assume 2.14 would actually mean GNOME 2.14 :(

live.gnome.org also contains RoadMap items on other pages :(

Example (probably created after this blog post):
http://live.gnome.org/Gnome_20Arch_20Progress

Another (links to some ideas):
http://live.gnome.org/ResearchAndDevelopment

Plus each product usually has its own RoadMap somewhere.

If someone could at least link these up...

Slobodan D. Sredojevic said...

Hello John,

I like your ideas or parts of them, especially call to "null" bugzilla counter :)

- Your remarks about "freezing UIs" are correct. I have experienced that with Nautilus when copying files from cds and Evolution like everybody else. There are more for sure. Can you please fill bugs for those you have encountered and open one META bug those filled will DEPEND on? That way when one is fixed others can be fixed too using the same/similar pattern.

- Like said in one comment - GNOME has appropriate release cycle, features and new desktop modules are added carefully (so far) (apple do this way also) but GNOME Foundation should react on quality on some modules, maybe they should feature-freeze some and call maintainers to fix bugs for cycle or two. Just look at weekly bug summary table at bugzilla.gnome.org for those problematic modules! And leave other modules to improve beside bug fixing.

- Not listing detailed feature list at live.gnome.org for 2.16 and beyond is marketing problem too. KDE has this and everyone know what to expect from upcoming releases. GNOME Foundation, GNOME architects, where are you going? :) Will a jabber client become part of desktop offer any time soon? This could be done long, long time ago, why not, why did you miss to do this?

Aldo "xoen" Giambelluca said...

I think one of the bigger GNOME problem is the website, it should be better, more updated.

I think also the GNOME platform is almost OK, but it needs some applications, and I think there are too applications that do almost the same work, and it miss some important features.

OK, choose is important, but too choice is not so good ;) IMHO the normal user needs some reference apps.

Many users needs :
* An easy way of set-up an internet connection
* Bluetooth (wireless) facilities
* A better control-panel/way of set-up some things
* A great buring app (with the possibility of backup DVD-Video), at the moment I use GnomeBaker, but it can be really improved
* A video editor (I know exists PiTiVi, Diva, etc...)
* A RAD, something like MonoDevelop but with a GUI designer, I know someone is working on the integration of GLADE into MD (This will increase the number of developers and it will make easier develop something)
* An app to manage packages graphically and over all distros (I know this is a little visionary ;))
* Eye candies (users like eye candy, and sometimes use something just because it's more beautiful)
* ONE music manager clean like rhythmbox/banshee and with almost all the features of amarok
* An IM client video-conference-enabled that can be used also with MSN Messenger Yahoo! Messenger and so on...
* More beer
* A free CAD program

Sorry for the brainstorming, for the moment I have a brain-buffer-underrun, but I think would be good for all have these things, expecially the beer ;)

Aldo "xoen" Giambelluca said...

I know is a strange place where to ask for it...but, is it possible the usual "official" pre-release tour of Davyd?

Here the previous tours :
http://www.gnome.org/~davyd/gnome-2-12/
http://www.gnome.org/~davyd/gnome-2-10/
http://www.gnome.org/~davyd/gnome-2-8/