[Clipart] Is anyone working on categorizing the existing images?

Bruno Coudoin bruno.coudoin at free.fr
Tue Jun 8 23:15:24 PDT 2004


Le mer 09/06/2004 à 07:26, Bryce Harrington a écrit :
> 
> > As for the choice between strong and weak categorisation, i would prefer 
> > weak. The way I see it is that categories are just keywords. This let 
> > the user create a 'package' of images based on keywords, and cross 
> > strong category boundaries. If I search farm, I'd like to see 'farm 
> > animals', 'farm tools' and 'farm houses'. Weak category is also much 
> > easier to administer, you need no central authority.
> 
> Agreed.
> 
> > One weakness of the database, as you propose here, is that you cannot 
> > install it easily and distribute it with a GNU/Linux distrib.
> 
> Oh?  Why not?
> 
cliparts are usefull for desktop users and for desktop users, as far as
I know, distribs do not install databases (like mysql) with the desktop
profile and that makes sense.

If you do a package with your cliparts, you will need to detect which
database is on the system, or force the use of one, install the package
and fill up the database. I am not aware of any package that does this
automaticaly, even if it's technically possible, it's rather complex and
error prone. Usually, there is a post step in the installed tool to
create the table and fill up the initial content.

Then, later on, there is still issues. In my application X, how will it
knows it can find cliparts in this database ?

You see, installing a bunch of XML files in a predefined directory is
much much easier.

Bruno.





More information about the clipart mailing list