Lists Home |
Date Index |
> Well, despite being a vociferous defender of UniodeWithBrackets, I can
> see the necessity for the infoset, merely for the community of spec
> writers. If you're describing something like schema or xpath or
> whatever, it's just way easier to do it in terms of the data model than
> in terms of syntax.
The Topic Map specification people have really taken the infoset to heart.
They have created a topic map infoset modeled after the xml infoset. They
use it to describe abstract processing of a topic map while it is
deserialized, in the hopes of obtaining maximum interoperability. This is
being built right into the new XTM specification, which formerly was only
about the XML syntax.
I find that is helps to clarify what the various components of the XTM file
are supposed to convey, and that is a good thing.
But then, Topic Maps need a standard processing model to work well, and XML
itself is (and should remain) much more general.