Re: GNOME 3.2 ideas and plans

[Martyn Russell <martyn lanedo com>]

On 22/03/11 19:19, Bastien Nocera wrote:
> On Tue, 2011-03-22 at 19:11 +0000, Martyn Russell wrote:
> > Forgive me, what's the difference between this and browsing images from
> > Tracker which mines data from Flickr already?
>
> Grilo is made to browse media, tracker isn't.

What are you basing that on?

Also expand on what you mean here by "Browse" please. As I interpret it, 
tracker can browse media right now and looking at the Grilo API, I see a 
lot of things that have been in Tracker for a while already and more.

> > I admit, the authentication part we do in Tracker would be better served
> > if it wasn't our own solution and instead using the "Broadcast" accounts
> > (as Ubuntu labels them) - if that's possible? Is this using libsocialweb?
>
> Ubuntu's "Broadcast accounts" is Gwibber's preferences panel. It's not
> integrated into the desktop. Tracker using libsocialweb for sign-on
> would be a step in the right direction, but I still don't see why it
> should index/mine websites I use to be honest.

Authentication to the side for a moment.

So if Tracker/Grilo shouldn't mine websites, how should they get the 
data for GNOME applications like shotwell, etc to browse image sources?

I should add, we're using official APIs from sites like Flickr.

The only way I see Tracker/Grilo being different is if Grilo does 
real-time searching with no caching of the results it gets. For web 
based services, this isn't going to be fast or responsive as using 
cached data. You also can't perform fast, complex queries that way 
(which modern apps and embedded platforms want these days).

--
Regards,
Martyn