Re: [entity-resolution-comment] Some queries on the XML Catalogsspec

[Richard Tobin <richard@cogsci.ed.ac.uk>]

> | Step 3 of 7.1.2 does not mention that the rewriteSystem entry with the
> | longest systemIdStartString is the only one used.  This is stated in
> | 6.5.5, but it should be explicit in the description of the algorithm.
> 
> Fixed.

... and the same for rewriteURI in 7.2.2.


> | Is there a recommended way to handle external identifiers when parsing
> | catalog files themselves?  Obviously they cannot be looked up in the
> | catalog.  Is it reasonable to have a copy of catalog.dtd built-in to
> | the parser and special-case its system and public identifiers?
> 
> This is a *really* good point and one that I stumbled over just
> recently myself.
> 
> (Lauren, I'd like to discuss this at the next meeting.)
> 
> Richard, what do you think the right answer is?

For RXP, I use a special entity resolver while reading catalog files.
If an entity has either the system ID or the public ID of the catalog
DTD, it uses a compiled-in copy of the DTD, otherwise it just uses the
system ID as a URI in the normal way without any catalog processing.
I suppose a schema processor that uses a catalog to fetch schemas
might do the same sort of thing.

Incidentally, this raises the question of copyright and licensing for
the DTD and schemas.  It should be possible to distribute copies of
them as part of software under any reasonable licence.

If anyone wants to try the RXP implementation you can get the source
from

  ftp://ftp.cogsci.ed.ac.uk/pub/richard/rxp-1.4.0pre4.tar.gz

Set the environment variable XML_CATALOG_FILES to a space-separated list
of catalog files, and optionally set XML_CATALOG_PREFER to "public" or
"system" (default is "system").

-- Richard