- From: David G. Durand <dgd@cs.bu.edu>
- Date: Wed, 11 Jun 1997 21:15:07 -0500
- To: w3c-sgml-wg@w3.org
>At 10:39 11/6/97 -0700, Andrew Layman wrote: >The example that was being discussed, <date>19960527</date>, clearly showed >that what is important is knowing what something is. Unless you understand >the notation used for this date you cannot interpret it. As I said in my >earlier messages, as we cannot use notation for XML elements, and we do not >have access to the lexical typing mechanism provided in the SGML Extended >Facilities Annex, we need some way to distinguish between the following >valid dates: > ><date country=US>01022001</date> ><date country=UK>01022001</date> (30 days later!) ><date country=Isreal culture=Hebrew>20010101</date> (This isn't even in the >same millenium!) ><date country=Isreal culture=Arabic>10022010</date>(This isn't the same as >any of the other dates.) In tradtional SGML, we use attributes and depend on applications to interpret them. We don't have a standard way to do this (which means that you may have to write date-checking code to verify your conforming SGML documents). Leaving things like this is _not_ a showstopper flaw. Most stylesheets need not know the data type of element contents, and most applications that do will require much more structure than a single element, and will use the doctype anyway (though they may have the DTD (and other data formats) compiled in). ><date>01022001</date> is clearly not sufficient to undertand the contents. >You need some qualifier (as I have been forced to do with application >specific attributes in the preceding examples). And it's not something that we desperately need, but something that we would like for some applications. >There are many other cases where you need to invoke some outside interpreter >to be able to understand what an element represents. exactly, and currently it's a processing decision -- if you're not doing _any_ procressing, then syntactic restrictions on the values don't really matter. >Either you need a >notation processor as per SGML or you need some other indirectable mechanism >for identifying how to interpret the data. stylesheet. >What I was suggesting was that >this should be done with a standardized attribute, which I called behaviour >for want of a better name. I do not believe this is about presentation, it >is about being able to make sense of the data when it is not in the same >textual format as the rest of the document. Its about understanding what >something actually is. Sounds like a great idea for a set of architectural forms that might be used by people with similar problems. -- David _________________________________________ David Durand dgd@cs.bu.edu \ david@dynamicDiagrams.com Boston University Computer Science \ Sr. Analyst http://www.cs.bu.edu/students/grads/dgd/ \ Dynamic Diagrams --------------------------------------------\ http://dynamicDiagrams.com/ MAPA: mapping for the WWW \__________________________
Received on Wednesday, 11 June 1997 21:19:17 UTC