>Len (to Bob G)
 

>From <p> to 
<partno>, information is gained.  In the other direction, it is 

>lost.   How does your 
continuum hypothesis work with that?
 
I don't see this continuum as a 
system that has to have forward and reverse gears, but as a model (as with treating time as the 
fourth dimension) to help determine other things, like appropriate choice of 
tools.  But first:

 

>Bob Glushko wrote:
 
>On one end are pure narrative things and on the other end are purely 
transactional ones
 
I've been using the terms 
"content-oriented" and "transaction-oriented" myself (they're all documents! 
They're all data!), so I was happy to see Bob use the word "transactional" here. 
Sean's quantitive approach may have 
more of a hard-science appeal than 
heuristics we might throw out (text nodes and element nodes as siblings: 
content-oriented; data starting or ending up in a DBMS: transaction oriented; 
etc.).
 
Back to tool choice: I've been thinking more 
about a potential correspondence between this dichotomy and the XSD/RelaxNG 
dichotomy. Look at all the DBMS vendor representatives on the original XSD 
Working Group, not to mention the e-commerce types (including a former employee 
of Bob's). Rick Jelliffe recently predicted [1] that <!----> 
XML Schemas 1.1 would be "<!---->slightly more DBMS-oriented and 
even less publishing-oriented." RelaxNG, on the 
other hand, is favored by leading proponents of content 
publishing-oriented document types: DocBook, the TEI, and (the W3C's!) XHTML 2.0 
working group. Sean has noticed this correspondence as well [2].
 
Can the 
data-DBMS-transactional XML vs. document-content-publishing XML distinction 
inform decisions about which schema language, and 
hence tools, are more 
appropriate for a given system?
 
Bob
 
[1] 
http://www.oreillynet.com/pub/wlg/4624
[2] http://seanmcgrath.blogspot.com/archives/2004_04_18_seanmcgrath_archive.html#108258054931482097