Wed, 20 Apr 2005
Couple LCA notes:
This is just a very well run conference, it seems, particularly for speakers. We're treated well and tonight's dinner was no exception.
The marketing BOF was a success, from my point of view, at any rate. There was very little structure- just a review of some key points, shutting up the Usual Suspects (including myself), and getting new faces to throw out some ideas which people can act on. I'll upload notes to live.gnome.org tomorrow. In general, we need to figure out how to encourage high-signal communication from new faces without increasing noise or stomping too much on more experienced people. That works in the Real World, like LCA, because the barrier to entry is high. How we make that happen in the e-world, I'm not sure.
Before marketing, we talked a bit about three point zero. I'm more and more convinced that we're not going to get to 3.0 as an organization- we're too afraid to fail (for reasonable reasons), we have too few resources, we are too enamored with planning,etc., etc.- I'm now fairly convinced that 3.0 is most likely to happen when someone goes out, experiments, probably fails, but get people interested in their experiment (and maybe their failure) and gets momentum about finishing the experiment and turning it into reality. That experimentation will light a fire under GNOME's ass and encourage new blood as well.
It seems to me like there are a few top-level 3.0-leaning notions that could be explored by people in their spare time, basement-lab style:
- search-as-interface (including tagging, metadata creation, nautilus-as-vfolders, etc.)(aka storage-lite)
- top-level-objects as interface (re-orging the interface on projects, people, etc., instead of applications)
- collaboration-as-interface (provide APIs and re-imagine apps to see content creation as a fundamentally group-based activity)
- document-based interface (instant-apply, recent-files as dominant menu item, etc.)
All of these areas have overlap, and all can happen to a certain extent within the 2.x framework. My gut sense is that 3.0 will happen when someone takes 2.x, beats one of these things onto 2.x, and says 'here is what I have done, here is how to build/play with it, I think this will be 3.0'. Note that this is not when someone talks about doing it- we can't really test, play, and argue until there is code, nor will the casually interested really get involved until there is something they can build and fight with. (Possibly 4.x is when we get two or three of the four...)
I'd love for UI people to be designing this, and hacker teams to be tackling implementation in an organized fashion. That's the Best Way. But that's not happening; the only large coherent teams that exist are reasonably focused on more conservative, short-term needs. I'm becoming convinced that it is someone in a virtual garage somewhere that will actually make us act on 3.0 instead of continuing to discuss it endlessly- we may not adopt their exact solution, but it'll kick us (or someone more active who supercedes us) into action.
This looks and smells a lot like I'm begging for a fork. That's probably right, at least short-term. GNOME is too large and too conservative for radical change right now. Looking at it from a more positive perspective, GNOME and associated projects have put together a lot of the pieces for such change to happen- hacky but implemented versions of Storage could probably happen relatively quickly on top of Beagle (and be a discussion point) or collaboration stuff could take place on top of gossip+gnome office, for example. Someone having a vision and going and JFDIing (instead of talking about it on planet or desktop-devel or whatever) will take those pieces and go, I think, and that will be awesome.
[FWIW, the language discussion already looks a lot like this- formally, GNOME is still dithering and not acting, but the grounds for argument are being set by those writing the code, and some would argue (not unreasonably) that tomboy, beagle, muine, etc. have not just set the ground but have already won the battle.]