Re: Proposed module: tracker

[Emmanuele Bassi <ebassi gmail com>]

On Tue, 2007-01-09 at 09:09 +1300, John Stowers wrote:

> > I have strong objections to the inclusion of tracker into GNOME 2.18.
> >
> > Actually, I have one strong objection, that is: what's proposed for
> > inclusion and where?  If it's tracker-search-tool (the UI), I'd say that
> > for what tracker *does* right now, there's no difference between tracker
> > and gnome-search-tool in gnome-utils.
> 
> This is true only if you look at the two UIs. Tracker adds search by
> metadata, (including user defined metadata) such as tags. Not to
> mention the quality and speed of the search is a multitude better
> using tracker.

sure, since it does return the content of a small subset of the things
I'd like to search, I'm sure the speed and memory footprint will be
small.

> If what's proposed for inclusion
> > is tracker-the-indexer, then until we have a use for the indexer in more
> > than one application, I'd wait for its inclusion; same goes for
> > tracker-the-database.  I'd also like to see a tracker-library to access
> > the data without having to implement the D-Bus calls into each and every
> > application.
> 
> There is libtracker

which is a really thin layer around the D-Bus API - so thin it looks
like the autogenerated stuff from dbus-binding-tool.  I mean a real
library, giving me the facilities tracker allows me - as an application
developer - to use.

right now, libtracker is like the raw and low-level libdbus; where is
the GMainLoop integration?  Objects?  Signals?  Something a bit more
high level for people *not* writing tracker?

> > I'd also like for tracker to become less of a moving target: in the past
> > six months tracker changed the database backend twice (at least), API,
> > UI; and it still indexes just plain text files, images and audio files,
> > with all the interesting stuff (emails, contacts, im conversations,
> > bookmarks, etc.) marked as TODO.
> 
> Are backend changes less relevant when hidden behind a stable
> libtracker api?.

if the D-Bus API changes, if the reliability and feature set changes
this much in just six months, then tracker is *not* stable enough to do
serious development work into applications.  this is

>  Personally I would be ashamed if tracker was left our
> becuase of the amount of svn activity it has seen.

personally, I would be *incredibly* ashamed of rushing out some
half-complete API and library, possibly breaking it each time for
another couple of releases.

> In the last 12 months tracker has been one of the most active projects
> in GNOME, Jamie addressed a multitude of concerns by adding
> considerably stronger RDF-type-semantic-web features since the last
> time tracker was flamed into oblivion on ddl.

you don't need to be in the GNOME desktop platform to be "active" or
even "used"; but, if you enter in the desktop release, then you must
adhere to standard of quality that such a young project as tracker won't
be able to maintain.  libraries in the desktop release *can* break
API/ABI - but they should not, unless stricly necessary; is libtracker
stable enough to make this guarantee?

ciao,
 Emmanuele.

-- 
Emmanuele Bassi,  E: ebassi gmail com
W: http://www.emmanuelebassi.net
B: http://log.emmanuelebassi.net