Re: GNOME online desktop - (some of the possible) next steps


Jeff Waugh wrote:
<quote who="Havoc Pennington">

  - modify (to become to support web apps

  - make and work as OpenID providers

Quick Q about these so I have my facts straight: Do you intend for mugshot
to be a branded 'frontend' to Which data will be stored on
either side? Or will the users and data be separate and/or federated?

There is probably a short and a long term answer.

Right now what we're doing is just adding more jsps to that will be served when the server is reached from These jsps will not have the Stacker, just an "edit my account" page and a "login/signup" page. They will have GNOME branding instead of Mugshot.
Bryan is making mockups.

The backing database is still the same, so an account on is the same as one on and the data is in one big soup.

What this gives us is by the end of the week or so and BigBoard etc. can start to be powered by

Longer term we can split things up more fundamentally. The idea is that Mugshot is the "stacker"/"watch what's happening online" site and will be parallel to / on similar footing with other third party web sites we integrate with. will have /applications, storage for user settings, your avatar/name, and other important desktop backend features.

Basically will be required for online desktop (but not for GNOME, if you just stick to "classic desktop mode"), while will be the stacker app and you only use it if you want the stacker feature. But isn't "core" in any way, it's just another web site we integrate with - though an open source one.

It isn't 100% clear yet how the details work out but that's the general idea that makes sense to me.

The long-term location of the servers isn't clear yet, and we'll probably also want to make the servers a test case for a shiny new Free Terms of Service / Open Service Definition that isn't defined yet.

The two items you quoted from my post, in more detail, are:

 - right now we "package" some web apps like GMail in the
   online-desktop module, adding .desktop files and MIME handlers
   and so forth, but we don't track these web apps on So we should track them, probably based on
   browser title (but never sending the browser title to the server)

 - there are Java libraries for OpenID; we should make your profile page
   on both and act as an OpenID provider,
   so people can use these for OpenID if they want. Also, this would
   let us federate with something like the abicollab site, potentially,
   *waves hands wildly*


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]