Sequence <was> Access Languages ...
Mark L. Fussell
fussellm at alumni.caltech.edu
Sat Nov 22 21:38:00 GMT 1997
On Sun, 23 Nov 1997, Rick Jelliffe wrote:
> > From: Mark L. Fussell <fussellm at alumni.caltech.edu>
> > I am not saying sequence is unimportant, but I think SGML is overly
> > focused on it (from an IM perspective) because it comes from a
> > paper/linear background. Information is rarely linear: it is only time
> > that is, which has caused some media [and the humans who use them] to be
> > (mostly) linear also. It can be difficult to break that linear
> > assumption when it doesn't apply if your tools keep reinforcing it.
> But do you think HTML would have become a popular markup language if its DTD
> was like this?
> <!ELEMENT html
> ( h1*, h2*, h3*, p*, I*, table*, tr*, td*, th*)>
> This is reductio ad absurdum of what you are saying. A DTD where all
> sequence information is made explicit.
I certainly wasn't trying to say that sequencing should be removed but
just that it can be difficult to see when it doesn't apply. Sometimes
information is (at least dominantly) organized as a sequence: Ordered
Sections contain Ordered Paragraphs. Sometimes information does not
inherently need to be sequenced but the application would like it to be so
it does not need to worry about ordering it at presentation (I am thinking
of a list of citations where there is a natural ordering [by one of the
columns/attributes of a citation]). And a variation of this case is:
sometimes it is just easier for people to take direct control than to do
informational markup. I think HTML and Word Processors represent this end
of the spectrum.
> If people just want a database dump format for nice relational tables,
> comma-delimiter formats are available and attractive. But when they have
> text which they don't want to have desequenced, SGML/XML can be useful.
Well, I guess I have larger visions of what SGML/XML can do, and I think
it is within (or at most a mild extension) of the original vision.
Requoting [Goldfarb 90, A.2.40]:
Generalized markup is based on two novel postulates:
a) Markup should describe a document's structure and other attributes
rather than specify processing to be performed on it, as descriptive
markup need be done only once and will suffice for all future processing.
b) Markup should be rigorous so that the techniques available for
processing rigorously-defined objects like programs and databases can be
used for processing documents as well.
SGML is designed to describe information, and although the original vision
may have been focused on describing documents I believe that was just
because it was the particular task at hand.
> ... Just folks can write
> HTML, at a pinch. If you go too much to a database mentality, you move
> to requiring custom-tools for data entry, rather than simple text-editors.
No argument that HTML is easier for novices to directly write than more
structured information, but that also applies to any of the more
sophisticated DTDs. The benefit of a human-readable and
human-understandable encoding like SGML/XML is that people can progress
from simple DTDs like HTML to more complex ones and still understand what
is going on. I have done this with web-site development where content
writers now use a "real" DTD that allows generation of different HTML
views (and more sophisticated linking... etc.)
And I do agree that accurately modeled information (e.g. normalizing in a
RDB context) can make it too hard (for the desired writes) to enter data
directly. It is likely that some SGML/XML DTDs will be designed to
contain all the necessary information with explicitly desired redundancy
and artificial sequencing but with the assumption that the processing will
later remove them on the way to the information model. This is almost
exactly what UI Forms and relational views are doing.
So I don't want to get rid of sequence, I just believe people should think
twice about it and assertain whether it is really part of the information
and is the best way to represent that information.
mark.fussell at chimu.com
i ChiMu Corporation Architectures for Information
h M info at chimu.com Object-Oriented Information Systems
C u www.chimu.com Architecture, Frameworks, and Mentoring
xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)
More information about the Xml-dev