> Well, this is at the very beginning of the development still. The high
> level architecture uses client-server, so there is a backend and
> frontend. All media is cached using SQLite and filesystem changes are
> automatically updated to the cache with pyinotify. At the moment I have
> image, music, video and feed caches. Movie and TV-series metadata is
> downloaded automatically from IMDb. Also album art and song lyrics are
> downloaded. Feeds are updated every now and then (interval can be
> selected).
Cool. One thing I've noticed when dealign with such applications is that
projects appear to always reinvent the wheel. Amarok, Myth, etc. etc.
all have their database (SQLite, MySQL or whatever) that cache the
changes. I wrote about this a long time ago and about the need to
standardise on a common framework such that the indexes and caches could
be reused in other apps.
I Agree completely!
http://colin.guthr.ie/general/development/metalibrarian.html
Since writing that, I think something like Tracker
(http://www.gnome.org/projects/tracker/) could provide the necessary
backend indexes rather than doign it yourself. The trouble is, there
then becomes the problem of KDE Vs Gnome Vs Xfce etc.
Tracker was originally meant to be Desktop agnostic, but it doesn't seem
to help in terms of KDE stuff. Strigi
(http://www.kde-apps.org/content/show.php?content=40889) appears to be
the KDE 4 indexer of choice.
Where some hope lies is that there now appears to be a FDO standard for
querying such indexes: http://freedesktop.org/wiki/XesamAbout
I just hope that such a system will work. Perhaps it's worth looking at
for your project? Just my thoughts - it may not be practical at this
stage (of both your and other projects), but I hope this is where things
will go in the future.
There is already a Xesam client library (in python) at
http://www.grillbar.org/wordpress/?p=202 that you could use to get
files and file metadata
John
--
To unsubscribe send a mail to [EMAIL PROTECTED]
