Re: XTM Datatypes [Was: Adding weigths to associations]

Lars Marius Garshol

> (I take it this means that you are abandoning the requirement that XTM
> support the definition of custom syntaxes.)

There is a difference between having a requirement and having a hope. I have
abandoned hope of getting support for custom syntaxes, but not the
requirement for them. I believe this is where the revised RDF specification,
which now allows you to use generic identifiers as default predicates using
techniques similar to those used in HyTM for role types will start to beat
XTM hollow.

> Why do you need to flag the document element as being a topicMap
> element? Surely it's enough to be able to detect one such element
> within the document?

I meant the topic map element, which must be the root of the extracted XTM
topic map, not the root of the source. (And I wanted to allow there to be
more than "one such element" within a document!)

> | If I want to retain information about the type of each role so that
> | I can select views of the topic map based on roles I presume you
> | must expect them to be re-constructed as topics within XTM.
>
> Yes. Is that a problem?

Only in that I have no way of distinguishing topics that have been
constructed from roles from those that have been defined for other purposes.

Martin