Re: Module semi-proposal: gnome-shell

[Seif Lotfy <seif lotfy com>]

2009/11/2 Jamie McCracken <jamie mccrack googlemail com>

On Mon, 2009-11-02 at 17:12 -0500, Owen Taylor wrote:

>
>     My initial understanding of the Zeitgeist engine was that it was
>     a data collection engine to collect a rich view of how the user
>     used their computer over time, which would then be used to build
>     an OLPC style journal interface; but that understanding fuzzes
>     at the edges when people are pressed about this, things like
>     deducing related documents from temporal overlaps and tagging
>     enter into the picture. This doesn't make me comfortable.
>
>     There are also questions here of the relationship with Tracker. If
>     Tracker really lives up to its promise, shouldn't timeline
>     information simply be extra metadata added in the Tracker store;
>     after all, a timeline really is just an indexed and extended
>     view of the classic ctime/mtime/atime metadata?
>
>     If querying the Tracker database for this is a) not sufficiently
>     efficient b) too cumbersome to code c) requires expert training
>     in RDF, then that, to me, would throw doubt on the whole Tracker
>     enterprise.
>
>     What would make us most comfortable would be a comprehensive
>     picture of how Tracker, Zeitgeist, and Nautilus work together
>     with the shell to allow finding your stuff. Now it is probably
>     not completely realistic for me to hang await for this to show
>     up in my inbox in finished form, so the first step (from my
>     technical perspective) is to get a clear statement of what the
>     Zeitgeist engine does, what new user interfaces are enabled by
>     that operation, what it does *not* do, and how it relates to
>     Tracker.

I dont want to comment too much on zeitgeist but AFAIK all querying,
tagging and event logging can be done in the tracker DB. A tracker miner
could be created to perform such logging although it seems the zeitgeist
team want to use a python daemon instead.

Tracker has tried not to step on their toes and instead encouraged it to
use tracker as a backend which they are receptive too. (they are far
happier using tracker if it were included in Gnome). AFAIK, all their
queries can be expressed in sparql so tracker should be a good fit

Timeline data is not persistent metadata like title/subject/mtime. Its
what I would call semi-persistent in that you would probably want to
store it for a limited time only (user preference for 3/6/12 month of
history). This could be added to tracker to stop it bloating up over
time. However the main concern I have is how much logging is done.
Zeitgeist, from what I heard,  appears keen to log quite a lot of data
which would quickly bloat up any DB. I personally favour a coarse grain
approach that only logs important stuff like file, application and web
histories.

I personally would like to merge zeitgeist functionality into tracker
but really its up to the zeitgeist team on whether they are willing to
do this (it would mean converting the python code to Vala/Genie or C as
well)

jamie

_______________________________________________
desktop-devel-list mailing list
desktop-devel-list gnome org
http://mail.gnome.org/mailman/listinfo/desktop-devel-list

Hey,
Thanks alot Jamie for actually explaining most of the sutff I wanted to exlpain.
During the course of Zeitgeist development we got into contact with alot of Tracker devs.
To sum it down. Tracker and Zeitgeist are not the same. 

Tracker answers question relation the metadata as well as the content of data. "Get me songs from artist A", "Get me all docs with tag B"
Zeitgeist covers how the data is used. "Get me most used applications and documents within a time period", "Which files did I have open while editing X"

 It is our plan to actually push our metadata about the items into Tracker keeping the info about the events for ourselves. I patched a dataprovider for Zeitgeist to send events to zeitgeist and the subject of these events to tracker.
But also as discussed with some Tracker devs a total merge makes no sense. However a dependency is possible and also a current option that will be tested during the hackfest. 
We also log every focus switch to generate contextual relevancy. This would just bloat the the Tracker storage. We have an extra table to handle this.

I will write a better explanation in a new mail on the tracker ML.
Cheers
Seif