Sunday, February 12, 2006

GNOME 2.16: Polish, Polish, Polish

The feedback on my blog after my last post seems to make it clear that there is some support for the theme/direction/goal of the next release of GNOME to be something along the lines of:


  • Reduce memory usage
  • increase speed
  • pay attention to (i.e. fix) the most reported user-visible bugs
  • resolve crasher bugs


In other words, we pretty much have the feature-set we need, now we need to concentrate on making the Desktop Environment not get in the way of users as they go about their work or play. Of course, the boundary between DE and application suite is somewhat hazy in many people's minds (not least my own!).

What is the feeling in the developer community toward this goal? The comments I am getting seem to be from people who characteris themselves as users, although I am sure that some of them are developers also. I would really like to here from people about this.

P.S. Some people have been asking for a more permanent list of the points that I have been raising in these posts. That will be Coming Soon. ;-)

15 comments:

aldo "xoen" giambelluca said...

I think GNOME can be faster & lighter...

...But I think the desktop is not so perfect as you can think.

I think it should has Bluetooth/Wireless/Network/etc integrated, I know there are project like NetworkManager, GnomeBluetooth etc but these are not integrated!

I think also GNOME should depend more on GStreamer.

Faster is better, but the work isn't finished yet!

aldo "xoen" giambelluca said...

Ah and of course also GNOME Power Manager should be into GNOME!

Anonymous said...

As Steve Balmer so famously once (sort of) said:

Memory, memory, memory, memory, memory, memory, memory, memory, memory, memory, memory, memory, memory, memory, memory, memory, memory, memory, memory, memory, memory, memory, memory, memory, memory.

David Nielsen said...

Everything needs to be moved to using GStreamer, especially ekiga (which also needs a serious UI review).

I would like to see serious music management software like Rhythmbox introduced to GNOME, we all run it anyways, it might as well become a solid part of the platform.

J5' SessionServices work seems to have gone into the work was done then forgotten pile, very sad since this was extremely cool.

The Epiphany downloader needs to support Torrent downloads

Palm syncing needs to just work, currently it's utterly broken in every distribution out there.

We are far from there yet, but we are closer than ever. We are just hovering around aimlessly, we should be bold enough to set some concrete goals.

Like the 2.15 cycle will introduce such functionality or we want a certain performance and memory goal hit (this is especially important if we want any shot at becoming the desktop for the OLPC machines).

I would love to see GNOME 2.16 aim to be OLPC compliant, say we increase the cycle to 9 months and keep this one aim in mind to give us time to get good testing done. High performance and low memory usage, post good guides to introducing users to profiling so we can get good data. I, for one, would love to help out if it was made easy.

Anonymous said...

And not only palm syncing, there needs to be syncing with mobile devices, and non-palm laptops too!

Support and integrate the OpenSync framework. The KDE guys are already doing it.


But I do feel that memory, stability and polish should be the main theme for 2.16!

aldo "xoen" giambelluca said...

Hey maybe GNOME Powey Manager is already in GNOME 2.14!!

David Nielsen said...

Nope g-p-m was rejected for module inclusion. This is good though, Wonder boy coder Richard has plenty of nice ideas that can now be implemented.

Oh and I keep a blog here:
www.lovesunix.net/blog
Where I also comment on GNOME, in case anyone is interested.

Anonymous said...

> # Reduce memory usage
> # increase speed

Thank you for raising these, John!

Pēteris Krišjānis said...

I think we should rethink GNOME release policy:

* Polish & bughunt and solving, serious Q&A before such releases
(I suggest such releases to do once in three months), NOTHING additional in features. Such releases will help distros not to patch their GNOME packages for serious bugs and everything will be more connected.

* Developer releases once in six months (new features, serious UI fixes, etc). These releases are for community for testing out new things before they go official (such background changing dialog changes, etc.). This is just to test playground and therefore never thought for official use. Once feature get's accepted in this release trough user feedback, it get's into...

* GNOME year's release, tweaked and tested trough some 3 or 4 beta releases.

So in summary, I suggest one official release in year (get serious, we must decrease count of releases for common crowd), four bug fix and translation update releases trough year for this release, AND special developer branch which is actually last year's release + added features and bug fixes, which are too serious and big to be regular releases. Feedback will go trough distros/early adopters/testing trough Novell/Sun.

It is time to get serious. Yes, GNOME is free software project, and everyone can contribute and that stuff. However, people start to depend on us, so we should care about them.

What do you think? :)

Pēteris Krišjānis said...

David Nielsen: with Ekiga is a problem, author simply has no time for this in this moment, also it is somehow difficult, because gstreamer isn't actually sound server, but system designed for UNIX platforms, but Ekiga aims to have Win32 port too.

So it won't be easy. I know, Ekiga looks promising, but really showstopper bug is that Ubuntu and others uses ESD for software sound mixing.

Maybe someone will rise a bounty to have at least Gstreamer output/input support.

Anonymous said...

Maybe it is better not to go for a new release until the polish is reasonably finished. I (and other users probably too) associate with a new release always:

* new features
* new bugs

So, to upgrade is always somewhat of a risk and also a chance. To be honest, I am somewhat sick of frequent upgrades every 6 months, especially since I don't see anything new in the recent releases.

As very important, I would consider, that Gnome:

* defines clear developer interfaces, on both - binary and source level
* _document_ them clearly
* keep them stable in long term

So that third party developers have a reliable platform for application development. Third party developers are not only open source hackers... the Gnome developer platform should also attract more proprietary software companies.

And - please leave unnecessary applications out of Gnome (someone said, rhythmbox should be part of Gnome). Please, keep Gnome as small and modular as possible and include only things that are really necessary for a desktop environment. It is not the goal of a desktop environment to provide killer applications for every task that any particular user might be interrested in. The goal of a desktop environment is to provide an infrastructure to run, integrate and develop applications - and this is difficult enough on it's own.
If users want to carry out particular tasks, they can install whatever third party application is necessary or they like. It's sufficient for Gnome to have a small set of builtin tools, like text editor, calculator, system management tools... but keep the balance right.
KDE still ships this "all in one" thing... Desktop together with every application you can imagine - and that for I dislike KDE. It's especially annoying when you have to upgrade with a modem connection!
Gnome is already more modular, but to integrate more and more applications in the core is definitely a bad idea, IMO.

Anonymous said...

I can see a big future for GNOME on ultra-portables. whether you agree with the above or not, faster and much lighter is good in anybody's book. Lighter to me says leave Mono apps out! I don't want yet another set of dependancies and the overheads of virtualisation that come with Mono.

I'd love a non-mono alternative to Beagle and there seem to be a few indexers / search tools which do part of the job quite well. Similarly I'd like something like iFolders but why would I want to use an alternative database engine when SQLite is used by so many other apps?

I love the alternative frontends for Evolution Data Server that are springing up. (Contacts and Dates by Chris Lord). Sure, I want my data to work together, but I don't like the heavy outlook-style all in one interface of Evolution. Perversly (possibly) I would like better Calender / To-Do / Contact integration with my desktop.

GNOME desparately needs a killer World Clock / alarm / app/task sheduler (Although I read there's a new sheduling framework in the early stages of development, was that on freedesktop.org?).

I much prefer Abiword + Gnumeric over OpenOffice. Gnumeric is sweet but Abiword is pretty slow to start (Latest Ubuntu version). Can't wait for Criawips to come of age and find a new name.

Would love something like Freemind as a GTK+ app, although there seem to be a couple could be going that way. The good thing there would be to use it as an outliner for Abiword, or Planner (Which rocks).

Would be nice to be able to dock my toolbars from Big apps (GIMP, Abiword, Gnumeric etc. ) to a tabbed inspector style box a la Apple Pages (Shock!).

Nautilus could learn a couple of things from ROX, GBrowser and the PCMan File Manager. Love to see application directories implemented.

Viewer app more like Apple Preview would be great too.

So, not much of a list :-) Oh yeah, did I say faster and Lighter? ;-)

Newton Gonzalez said...

Hi:

I read your two articles published in the gnomejournal about "Marketing GNOME" and i can say that agree, in some things, but mabye not at all.

In my personal opinion gnome must have better documentation, maybe the hackers of the development are genius, but we the normal people yes we need.

And i guess that is necesary better tools for development (with better i mean easier,maybe of kind of delphi, or c++ builder, but for gnome), exist anjuta and glade, but them aren't so integrated and don't work properly at the first; maybe that kind of tools are not so necesary for the development of gnome core, but yes for create new aplicationes, in gnome are necesarry a lot more of applications, and not all can be created by the team of gnome, but can be made by the comunity.

must be polish, and with minus requirement of hardware, not all the people have newest machines, and if we want a desktop for all the people must be possible of running on that kind of machines.

Is neccesary know what the people want, how them began to use gnome, if them are tested other desktop, and a lot of things more that give to the team knowlodge of the market, and mainly of his user bases. I guess that can be achived trougth a agent that can make some kinds of polls periodically, polls shorts but with the finality of gatter valious info that would guide the future of gnome but wiht direct info of the users.

Anonymous said...

I think developer tools should be better, especially Glade (or Gazpacho). It should be easy to do your own widgets and use them in the UI designer tool.

CreditUser said...

As for me, GNOME suits me completely. I use it often and I am really satisfied.