On Fri, 7 Dec 2001, Steve Harris wrote:
> On Fri, Dec 07, 2001 at 10:28:01 +0100, Richard Guenther wrote:
> > > An ontology in other words. Good idea BTW.
> >
> > I even want you to consider a more generic approach for exposing extra
> > information. Add a (string) key/value array to the descriptor where you
> > can store such information (and be happy with the generic interface in
> ...
> > int nr_extras;
> > struct { char *key; char *value; } *extras;
>
> Actually that isn't quite general enough for what I was suggesting, you
> need:
>
> int nr_extras;
> struct { char *subject; char *predicate; char *object; } *extras;
>
> Because you aren't neccesarily asserting things about the plugin, you
> might be asserting about the categories.
>
> For example you could say that "this plugin" is a type of "state variable
> filter" and "state variable filter" is a subclass of "filter". If "filter"
> was a predefined category, but "state variable filter" wasn't.
Yeah, this would be even more flexible. I like it a lot.
> But anyway, my feeling is that this stuff is best stored outside the
> plugins data.
Well - at least such stuff has to be "packaged" with a plugin, so
the plugins data would be a natural place to store (or at least to
access) it.
Richard.
--
Richard Guenther <[EMAIL PROTECTED]>
WWW: http://www.tat.physik.uni-tuebingen.de/~rguenth/
The GLAME Project: http://www.glame.de/
