🚀 go-pugleaf

In <1993Jan7.180136.236@infodev.cam.ac.uk> Alasdair Grant writes:
>
>Thanks for your reply.  I just want to pick up on something - what do
>you think about including features like comments, tag minimization,
>and data tagging in a database?  If you don't, you aren't really storing
>the "real SGML", but I honestly can't see the point of remembering that
>tag of item X was minimized while tag of item Y wasn't.  Seems to me
>that there is at least one level of SGML which gets in the way of
>database applications, and ought to be stripped out in any importation
>of SGML into a database.
>

It's certainly technically possible to preserve all of that in
a database if the needs of your system require it, but I wouldn't
loose sleep over a system that didn't preserve comments.  Tag
minimization can be re-applied algorithmically since the minimization
rules are defined in the DTD, and features like shorttag and
datatag are best left alone, since they are primarily concesions
to the problems of data entry, not information identification or
structuring, and therefore shouldn't enter into the equation.

Most of the folks I've talked to do not seem to consider the requirement
that SGML source be normalized for database storage to be a problem
in practice, and we know we can unnormalize it if we ever want to.
You can get around the problem of preserving comments by providing
elements whose defined semantic is to contain descriptive comments
about the elements that contain them or define explicitly-descriptive
elements for the sorts of things you normally use comments for,
such as "file headers" that contain author information, change
tracking info, copyrights, etc.  That is the approach we've taken
within IBM.  I will not accept any arguments for preserving comments
based on the needs of applications that use the data in comments
for their own purposes.  SGML provides better places for such
information, such as processing instructions, the APPINFO statement
in SGML declarations, extra-SGML mechanisms unique to the
application, and so on.  Since comments are, by definition, not
part of the document content, and are not carried by the ESIS
information, there is no strict requirement to preserve them.

Eliot Kimber                      Internet:  drmacro@ralvm13.vnet.ibm.com
Dept E14/B500                     IBMMAIL:   USIB2DK9@IBMMAIL
Network Programs Information Development     Phone: 1-919-543-7091
IBM Corporation
Research Triangle Park, NC 27709