Decadence of the Enterprise Desktop
I’ve been following the “Decadence Thread” on Planet Gnome with great interest. It all started with Andy Wingo’s blog post “gnome in the age of decadence”.
The blog posts are getting harder to track so for my own and hopefully other’s interest I’ve written down a resumé of what has been said so far. Note that the resumes are totally my own perception of the individual blog posts and to pay respect to this discussion you should really read the originals.
Here, newest at the top, earlier at the bottom.
- Luis Villa – (new) Has talked about this subject before and drops a handful still-relevant links to his old posts.
- Havoc Pennington – (new) By sticking to the desktop metaphor we will not innovate or gain significant amounts of new users. We must know the audience which we try to address. Do not look at old Gnome 2.0 heroes for inspiration or confirmation.
- Rodney Dawes – Thinks that we should not trap our (and potential user’s) minds into a “Desktop Metaphor”, but focus on empowering users.
- Richard Hughes – Thinks that it is nonsensical to argue that we need a Gnome 3.0 to develop big new features. He emphasizes that people “use GNOME to actually do things”, not to look at wiffy effects. We should not develop fancy effects just because we can.
- Jono Bacon – His own words “predictable and reliable, but has ceased to be exciting and innovative”. Points out there are a lot of people hashing out visions and new ideas, but very few backing that up with real code. Gnome 2.0 owes a lot to our Rock-Star Coders and we need buy-in from them for anything to take off. A way to get something going was to put a select handful of our finest masterminds into a room, and when the white smoke emerges we have a battle plan.
- Thomas Thurman – Does not want change for the sake of change. Does not like cold meals because of broken applications and would like apps to asymptotically approach perfection.
- Alberto Ruiz – Wants closer integration with the OS and not just let Gnome sit on top of it. Instead of us adapting to each and every OS, Gnome should supply its own interfaces for integration, taking PackageKit as a good example here.
- LinuxHater – (new, warning explicit content
) Thinks that all Gnome developers are a bunch of wankers. There are tons of things left to do for Gnome, like tackling all the enterprise features of the desktop. - Johannes Schmid – Thinks that it is the right time to start thinking about our next-gen desktop. We should focus more on writing code instead of just talking about visions. Effects should enhance user interaction, not just be flashy.
- Andy Wingo – Chimes back in with some concrete ideas (his initial post did not have any “solutions”). Mentions the possibility of a Gnome “skunkworks” where experimentation is encouraged. Talks about enhanced user interaction, use of GL, and talks about Pyro.
- Owen Williams – (new) (Only syndicated on Planet OLPC) Has a three point line-up of why we need to break API and ABI of gtk+ and argues for a Gnome sandbox project.
- Lucas Rocha – Belives that we reached the original goal for Gnome – a desktop that just works. Acknowledges that Gnome has some problems. Writes up a list of all the things we have achieved in the Gnome community and project as a whole.
- Christian Schaller – Points out that there have been a few whole hearted attempts at writing the code to implement some orignal ideas for a next-gen desktop. The current stable iterative process has given us great buy-in from distros and users.
- Calum Benson – Tends to agree with Andy. While Clutter is cool it will not solve our problem with “vague lack of coherence and integration on some parts of our desktop”.
Me? My opinion about this matter is that Gnome has indeed painted it self into a corner in some way. In the office we usually refer to Gnome colloquially as “The Enterprise Desktop” (a term I believe to be coined by the internet’s Luis Villa). It is very mature and stable, and indeed ready for the enterprise. But it is also seems to be mature and stable in a way that was more fitting several years back.
It is getting increasingly hard to get new stuff into Gnome, and when someone approaches with something that is slightly controversial huge flamewars erupt. Consider Tracker and Empathy on the desktop-devel list and the recent “incident” on the gtk-devel list.
On one side we have distros with very high bars for stability, and the others we have a very grumpy but vocal minority of users and developers who like to bash on anything unknown. Tight spot.
On top of technical problems with outdated semi-deprecated libs, and missing core funtionality, we also have a community issue. Sure “Gnome is people” and we have a rocking community and all that, but I am also seeing growing internal tensions and frustrations turning into poison, more and more often. I, for one, have closed my laptop lid with a bad feeling too many times.
Solutions?
I really have a lot on my heart about this issue, but I am too tired to write down the details for now. I will write it up tomorrow. To drop a clue of my intentions I don’t care much about the visual or interaction parts of a renewed platform, more about making a modern, extensible, maintainable, and cohesive platform (buzzword count alarm goes off). Doing all sorts of fancy visual tricks should be a lot easier when we have a solid platform to do it on, if it is properly extensible we wouldn’t have to break anything in the process. Anyway, I already posted some related thoughts a while ago.
UPDATE 1: I added Havoc Pennington’s and Owen Williams’ replies to the list. Somehow Havoc’s post did not show up in Liferea. I wonder how many PGO posts I’ve missed this way. Owen is interesting because he is not syndicated on PGO.
UPDATE 2: Added Luis Villa.
UPDATE 3: Add LinuxHater
June 11th, 2008 at 17:53
You missed Havoc and I think he spoted the problem when he says: “2) It’s broken because by definition if it’s still “the desktop,” it’s not a revolution. Everyone has a desktop already. What do they not have? Providing what they don’t have would be a revolution and could lead to 80% marketshare. ”
GNOME has a solution for almost anything in the desktop world as we know it nowadays and so we think (and we are right) that we are not innovating.
I think that the eye of the storm is on “Everyone has a desktop already”. If you want to innovate you need to do things that will not fit for the userbase, things like working with touchscreen UI or Internet tablet devices or phones/pda.
my 2 cents
June 11th, 2008 at 17:58
I really think the on-line desktop is the future. But I don’t see that as things like prism trying to wrap a web-app as a desktop app. I’d rather see the desktop infrastructure integrate nicely with the web. I’d like to be able to see my Google calendar with the gnome panel date applet, I’d like to be able to double click the date and open google (or yahoo, windows live, etc.) calendar instead of evolution, Then when I’m done editing my calendar I’d like to see it instantaneously pull the update from google. I’d like to be able to easily configure gnome to use various webmail providers as the mailto: handler. I’d like to be able to pull in e-mail addresses from social networking services into my evolution address book. (I admit this one is a little harder but the other two are almost trivial, there was a patch in bugzilla to unhardcode evolution from gnome-panel that was ignored!).
I also think to try to advocate free software gnome should collaborate (with KDE, Apache or who ever might be interested) in creating high quality free-software web services that can be integrated as another option.
June 11th, 2008 at 18:08
What Rodney said is true and we should keep an eye on this. Personally I don’t see “the desktop” going away anytime soon but we should be able to adapt to new media and devices. Our platform technology needs to change to allow this.
There are various interesting comments and different aims. Again, I don’t think “3.0″ has to be /radically/ different to the current offering but I’m sure we can do better than what we currently have, mostly in the core components (panel/applets/filemanager) and their relation.
Finally, a cleaned up and more powerful platform would make working on GNOME fun again I suppose,
June 11th, 2008 at 19:06
“Doing all sorts of fancy visual tricks should be a lot easier when we have a solid platform to do it on, if it is properly extensible we wouldn’t have to break anything in the process.”
I like your final thoughts, and will be interested to read a more elaborate post on them.
June 11th, 2008 at 19:20
Although I’m not syndicated on Planet Gnome, I am on Planet OLPC and I’ve written up some of my own thoughts on the subject. Could you include them in your list? http://ywwg.com/wordpress/?p=426
June 11th, 2008 at 20:15
I agree with you on that it’s very hard to add new features to gnome now.
One way to solve the problem is to have gnome-base derivative just like the distribution world. Someone package a Gamer flavor of Gnome, 3D Gnome, flashy Gnome, EduGnome… and let the market decide which one will win.
June 11th, 2008 at 20:19
Basically, in the free software world, when things getting staled, FORK it!
June 12th, 2008 at 00:37
A thoughtful post with a good resumé of the situation, of what was already said. At least, it seems that a lot of people agree on the fact that Gnome – or widely the desktop – should be “updated”. Gnome is a good desktop, corresponding to the metaphor of 5/6 years ago.
But, I disagree when you say that Gnome is enterprise ready. Yeah, Gnome is reliable and really usable, easily. But there some big points missing to be Enterprise ready: by example, a good Office suite and a good framework for synchronisation with mobile devices. Thooses services are used (heavily sometimes) in companies and we do not have something reliable.
I wait your post of today
June 12th, 2008 at 08:39
“Enterprise” is one of the last things that come to my mind when I think of average Gnome desktop. Lack of zero-configuration of things like email clients, poor kerberos support all around, poor administration utilities of kdc/ldap users, no acl support, poor tools for log file management, poor tools for reporting problems, poor support for enterprise IM (zeroconf is _must_), …
June 12th, 2008 at 19:27
GNOME should be a specification, so that many tools can conform to this specification and be a GNOME desktop.
GNOME apps should also require apps to separate the UI from the core (PCMEF/MVC/THREE TIER). Building all these apps without a core backend, is just bad, extremely bad software engineering.
By separating the core from the UI we could have very tight communication between the different module expected in a GNOME environment and the desktop will come alive and breath.
Too many “GNOME apps” has a bad architecture.
Put the N back in GNOME
June 13th, 2008 at 00:15
Thanks for this nice resumé!
My own thoughts on this:
http://thorwil.wordpress.com/2008/06/12/talking-about-gnome/
June 13th, 2008 at 04:55
Dear Gnome,
1) Make friends with Firefox and their toolkit and make it so that I can run more and more applications online. Eg, home desktop that syncs up data with a website, and a Firefox addon called Remote Gnome that installs Gnome XUL apps and lets me access that data. When on my home desktop it syncs things up. Gmail isn’t open source and there is no good open source equivalent — you need to fill the niche and help bring windows users over.
2) Don’t lose the perceived stability of Gnome. Ubuntu only got mainstream good as of 7.10 (IMO) and we need a few years of consistency to help people move over. If you announce any radical vision market it as a separate thing that doesn’t dilute focus (even if, in practice, it does).
June 13th, 2008 at 14:16
Hi,
it is interesting to hear about tracker and emphaty. Two projects I belive would enrich the GNOME platfom a lot.
1. Imagine complete menu powered by tracker, no delay when clicking Application, no delay when reading icons. Much more functionality.
2. IM and messaging as an integral part of GNOME and applications using it. Sharing taken to the next level.
3. Clutter – really nice animations and effects for the user!
4. Pulse Audio as Lennart recently showed there is a lot of space for improvement.
5. Continual improvement of all parts to work nicely together.
BTW, can you point to threads were tracker and epiphany was put down?
June 16th, 2008 at 12:59
Troll, I totally agree with you.