Re: Public Identifiers, and CATALOGS
At 11:18 PM -0500 4/1/97, Paul Prescod wrote:
First, I want to thank Paul lfor his "rant". Very nice. I'd also like to
acknowledge that (as he said in private email) PUBLIC IDs may not always be
bound to an immutable reply: resources like "today's weather report" change
regularly. So a caching application that inferred PUBLIC->SYSTEM mappings
from douments that it retrieved over the web would also have to track the
dates of validity on the fetched resource -- but, at least for HTTP, that
mechanism already exists.
>> This is not to say that indirectioon shuld not be supported, but that
>> if you provide a fallback URL, it will be the one you could have provided
>> in the first place, and fetching the document would have been faster
>> (no need to connect for CATALOG).
>David has described how getting the catalog can speed up your download
Well, actually the TCP efficiency may do that, but even more important to
my mind, is the fact it can help increase the ability to do incremental
display, by giving you the stylesheet before you commit to transferring the
document. Those of us sitting on a modem all day really live and die by
>> In that case, I hope we don't have to admit in public that CATALOG is
>> an ascii non-XML non-SGML file because the SGML vendors said it would
>> be too hard to implement if it was in SGML. Ooops.
>There is a difference between "too hard to implement" and "not worth the
>extra typing for a purely political point." But I agree that we should
>at least slap a DOCTYPE on in the same way that we do for simple DSSSL
>stylesheets so that the document type is self-describing.
I think that noting that the (trivial) format difference is determined by
compatibility should serve fine... We've already done that in several
places, with considerably more complex syntactic oddities, and it's not
like CATALOG is a punishing and tricky format.
>> Or worse, because if the PUBLIC ID works, you don't need to test the
>Sounds like an opportunity for XML tools vendors. The exact same problem
>exists today with URLs that work on your machine but not on the server
>for any of a variety of reasons.
Exactly -- if the requirements on authors to use PUBLIC IDs were hard, and
if URLs actually gave us guaranteed resolvability, this whole line of
argument would muster some conviction.
Everyone who has used a web-server on a case-insensitive file system and
then moved to Unix, has discovered that URLs are not as simple as they look
(even with the simplest of server and client implementations).
David Durand firstname.lastname@example.org \ david@dynamicDiagrams.com
Boston University Computer Science \ Sr. Analyst
http://www.cs.bu.edu/students/grads/dgd/ \ Dynamic Diagrams
MAPA: mapping for the WWW \__________________________