Consistency and Formal Model Re: [topicmapmail] Can a resourceRef be a topic

[Lars Marius Garshol]

* Sam Hunting
| 
| Sure. And if topic maps are to be a neutral envelope for the
| knowledge interchange, then it would seem a good rule of thumb to
| put as few constraints as possible in the topic map model, as
| opposed to applications of that model.

I am inclined in this direction as well. There are a lot of
constraints one might be tempted to put in, and quite a few that are
in there already. Having as few as possible seems better to me,
however.

| It's a lot easier to add constraints to the model than it is to take
| them out later, once they have been added.

Actually, it's the other way around. When you lift a restriction, all
existing data is still acceptable, so that is a backwards-compatible
change. Adding a restriction, however, is different, as it may
invalidate some existing data.

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
ISO SC34/WG3, OASIS GeoLang TC        <URL: http://www.garshol.priv.no >