[sc34wg3] Comments on Tau model
Sat, 28 Aug 2004 09:45:24 +0200
<this time me dancing on the table> :-)
Robert Barta wrote:
> On Fri, Aug 27, 2004 at 09:32:08PM +0200, Jan Algermissen wrote:
> > > > Ah! That's what you understand Topic Maps (as a paradigm, not the docs)
> > > > 'are'. Unfortunately, I have no idea what you mean by
> > > >
> > > > - infrastructure
> > > >
> > > > - information model
> > > >
> > Well, fair enough....OTH, you *do* have an understanding and appearantly
> > /tau satisfies it. Interestingly, it does not satisfy *my* understanding,
> > so I would be soooo happy if you could define/explain the terms (loosely).
> > I PROMISE, I won't do nitpicking on the language :o)
> OK, I have alerted my lawyer, just in case :-)
> We need languages to write, modify, transform, constrain
> and query topic maps. These languages (preferable with a common
> set of concepts) is the 'infrastructure' I am talking about.
Would you agree, that this is a secondary issue, since the infrastructure
will just be defined in terms of the (using your term here) information model?
(Like SQL is defined in terms of relational algebra)
> information model:
> I think I already covered it in an earlier posting, but my view is
> that at the moment we are dealing with the following information models
> - table-like (based on tables with cross-references)
> - tree-like (earlier hierarchical data models, now XML)
> - graph-like (RDF, TMs, ...)
> - and unstructured text streams
> These are first data models. Depending on your data, you may
> choose one or more of them.
> And then you have the technology: RDBs for relational data, XML DBs
> (or relational databases) for tree-like data, etc. This is more about
> speed and flexibility (and, uhm, cost).
> Hope this answers this a bit.
Yes it does and it makes me extremely happy, since this is EXACTLY my
approach to the whole story. Not sure when I will find the time, but
I should send a lot of follow ups on the above. Hopefully....
Just to make sure that I understand you correctly, by 'information model',
do you mean
- the structural abstraction of a data model (e.g. tables, trees...)
- the complete thing (information model == data model), including
structural abstraction operators, constraint-mechanisms
BTW, you might find this interesting in the context of the whole
comparision between rel. model, XML, Topic Maps:
PS: Do you have a new version of /tau that you could send me?
Jan Algermissen http://www.topicmapping.com
Consultant & Programmer http://www.gooseworks.org