Re: GNOME: lack of strategic roadmap

[Ivan Frade <ivan frade gmail com>]

Hi,

 The big idea behind GNOME3 can be to offer a completely new User Experience. GNOME2 did well with the usual Menus/panel/folders approach, it brought stability, performance and we built the basic blocks of a Desktop. Now comes the time to use those blocks to revamp how the user interact with its computer. In some way, it should do on the Desktop what maemo5 did on a mobile device: use the things we already know in a new way and it doesn't look like the usual boring computer.

 To achieve this, we have 3 big new components:

1) GNOME-Shell: maybe it needs a lot of usability testing, maybe it is polemic (which is good), maybe it needs more effort polishing the implementation, but definitely looks fresh. It is a new way to interact with the desktop.

2) Tracker: It allows as to have a central repository of information. Contacts, Messaging, Tags, etc. in one place. Applications share data, and specially LINK data from different sources. The user receive a picture from his contact X, telepathy saves that relation in tracker, and later we can show it in some fancy way, maybe relaunching the old Dashboard project. The user can set tags on contacts, documents, songs, pictures and see all that content together in the same view (maybe some FUSE or nautlius magic is needed here)

3) Zeitgeist: At the moment the Journal is the big use-case for Zeitgeist, and it is enough if it works fine. Functionality like showing "related contents" or "related people" when the user opens a document would be great and will come at some point.

There are still some open issues about how this components work together: how to show the things Tracker stores in a meaningful way in GNOME-Shell, or how to combine the information from Tracker and Zeitgeist. But GNOME3(.0) is just the first step, during the GNOME3 series we can do a lot of improvement a bring new applications and new ideas with these ingredients. Maybe during those releases we can try a tag-oriented nautilus, or a content-oriented Open/Save dialog. If we feel really brave, we can play with the idea of hide completely the filesystem to the user, why not?

Besides, I see some interesting features appearing here and there, and should be working fine in GNOME3 out-of-the-box. I would expect to start my brand-new computer at home, install GNOME3, and see all my media content in my PS3. We have the pieces to do that (UPnP support with Rygel). Then i take my laptop, travel somewhere, and the pairing with my phone to use its data connection should work fine with a nice and easy configuration dialog. AFAIK, we also have the pieces in place (NetworkManager). The user can configure its flickr/facebook/gmail account in one place, and the sharing is enabled in the applications with 0 configuration (maemo does something like that already). The "Network" part of GNOME finally appears :)

These are just few use-cases that came to my mind. I am sure some people out there can come with newer and more original ideas. We have also some other technologies there, waiting to be used in something valuable for the user, like ubuntu-one/coachDB, the epiphany-gtk thing that can show widgets embedded in the HTML, and _javascript_ is floating around without a clear place in the platform (and it looks more a more important everywhere). And as a side note, some not-so-new technologies are IMHO understimated, like avahi; and internally in the platform Vala and gobject-introspection can open new opportunities.

I don't want to reopen the (eternal) toolkit discussion. All these ideas can be implemented with the current toolkit. We will find the concrete limitations on our way, and if we need something more or something else, then we can think in a new toolkit, reuse another, move to HTML+_javascript_, move to silverlight... whatever, but we can still do a lot of things with what we have, so don't let it block other innovations.

Finally, to elaborate a roadmap, we could go through all the GNOME (Desktop) applications and check how can they integrate better with the Shell/tracker/zeitgeist, probably offering some new surprising functionality.

This is my euro to the strategic roadmap. I hope some people share the enthusiasm for offering something new in GNOME3, even if it is completely different from what i propose here.

Regards,

Ivan