[XESAM] Minutes of meeting 2007-05-15

[jamie]

On Mon, 2007-05-21 at 14:56 -0400, Joe Shaw wrote:
> Hi,
> 
> On 5/21/07, Evgeny Egorochkin <phreedom.stdin at gmail.com> wrote:
> > Not sure what you mean by "knowing how to display",
> 
> Simply that code has to exist which pulls specific information
> provided and displays it in an optimal way.  For instance, displaying
> some sort of widget for a "person" rather than a URI.
> 
> It must also know how to filter information so that a user isn't
> overwhelmed.  For a person, inundating the user with all possible
> information about that person in little more than a key-value format
> isn't useful.  Otherwise we might as well just be displaying the FOAF
> file itself and not bothering to parse it. :)
> 
> > but when app encounters meta-data outside of the core or known onto, it has
> > to deal with it nevertheless.
> 
> Beagle's policy has always been to drop this information, actually.
> If we can't present it sanely to the user, it probably won't be useful
> to them.  (There is also the conscious decision that Beagle is just a
> search tool -- it largely provides information "at a glance" to the
> user; if they want details, another application is better suited than
> we could ever be.)
> 
> > The more structured the onto is, the more opportunities the app
> > has to present/process the data properly.
> 
> Sure, but doesn't the app have to know about this?


not really. Tracker is moving away from hard coding stuff like this so
that it can handle custom services with a tile based gui without a
priori knowledge or any hardcoding.

EG see our service def files (note the TileMetadata fields!)
http://svn.gnome.org/viewcvs/tracker/trunk/data/services/default.service?view=markup

jamie