Another possibility is RDF.
This isn't to knock XTM -I think both are good solutions to your needs.
I mention it because there is a demo you can browse showing facet
navigation.

http://www.swed.org.uk/

This gets round the limitation you mention - there are organisations
that are both 'commercial' and 'not for profit' (part of the
Organisation Type taxonomy) - eg ACRE which is both a Ltd Company and a
Registered Charity.

One of the other things to say is that your ontology for 'location' and
'work area' (however you express them) may well be useful resources in
their own right. Or you may wish to search for existing taxonomies to
reuse. Good luck!

Cheers

Steve


Lars Marius Garshol wrote:

>* fojbe@...
>|
>| Browsing through job offerings can be by job location(s), job work
>| area(s) and salary.
>|
>| As I found out: A resource can not map to two or more headings in
>| the same taxonomy.
>
>You are correct about this restriction in XFML. I am not aware of any
>workaround for this within XFML 1.0.
>
>| If so, it is impossible to say that one job takes place in two
>| different locations and job description can not be in two or more
>| predefined work areas (say "java programmer" and "information
>| architect").
>|
>| I just started exploring things in this area, so maybe I am
>| conceptually wrong.
>|
>| What do you say?
>
>Well, one obvious alternative (to me, anyway) is to drop XFML and
>instead use something that does let you do this. One candidate would
>be XTM, which could do this easily:
>
> - topics (of type location) for locations,
>
> - locations have names, maybe finer-grained types like (city,
> country, state, ...), and maybe associations building a hierarchy
> (contained-in),
>
> - work areas you can treat the same way,
>
> - job offers become topics of type job-offer, with associations to
> location(s) and area(s) and an occurrence for the salary range.
>
>That should do it, but you can of course go further and create a more
>complex ontology for your domain if it seems useful. Creating a
>faceted search system for this ontology should not be hard. (I could
>show you an example of this on a public web site, but unfortunately it
>is in Norwegian.)
>
>
>