Re: Fwd: Shelling Epiphany

[Alexandre Mazari <scaroo gmail com>]

Hi Tirtha !


> Can this be taken up as a GSOC project? Is the scope large enough?
> >

Definitely any help would be greatly apreciated.
Actually this subject is part of the ideas proposed for GSOC:
http://live.gnome.org/SummerOfCode2011/Ideas
So it must be first accepted by the Gnome representatives, a mentor must
be found, an evaluation plan defined and code be written !

Regarding the scope, here is a rough overview of the tasks as I identify
them:
- getting comfy with gnome's, ephy and shell processes, people and
technologies (build system, tools, sources)
- define the interaction model (how the user creates/manages the apps,
launch them, remove them...). That might imply some dialog with devs and
designers before finding the accepted and shared view. Hard to estimate.
- implement a "webapp visor", whether inside ephy, as a simple view, or
as a separate tool (I prefer the second option, but debate is ongoing).
Separate cache, preferences and offline data is a plus.
- implement a "webapp creator", creating the right .desktop file,
fetching the favicon, letting the user choose the name of the app...
Again that could be done whether in an ephy extension or a separate
util. In that case, a small ephy extension should be created providing
the current page url to that tool. Here also, I favour the second
option.
- optionally, and controversially I must add, a bridge to the packages
system could be created. This way, the user could add/remove the webapps
using the standard pkg management tool of their distro of choice
(pkg-kit, synaptic...). The packages should be added to a user-local pkgdb
to avoid poluting the system-wide db and requiring admin credentials.
- work toward exposing system facilities to the web applications 
(Notifications, Icon Badges, FileAPI, preferences). This certainly imply some
work on webkit-gtk 
- allow web apps to be prefered apps (ex. Horde/GMail as default handler
for mailto: links)
- allow per-application custom script injection (see Fluid.app for OSX).
Those script might allow better system integration. For example, changing
the icon badges when the page title change for unread mails.
- an app "store"/repository of free software web apps (Identica...) ?
- push for the inclusion in the related components

All in all, 3-4 months seems a reasonable/tight timeframe for such an
endeavour, depending on the student current familiarity with the
technologies.