Lists Home |
Date Index |
> That's an interesting comment. I guess I would have made the seemingly
> opposite statement: "The SGML vision is exclusive". By that I mean that SGML
> was, and is, a pretty impenetrable morass to the uninitiated. When I
> purchased my copy of the SGML Handbook a few years back I found it a
> startling failure in communication.
Err, this is a comment on the SGML Handbook, not ISO 8879. They are not the
same. I find 8879 itself far more straightforward to use than in Charles' version.
People may be interested in a page count:
SGML - 85 pages (in my printed copy) without the annexes
XML - 52 pages (printed out from a web browser, at a medium to small setting)
Add a couple of pages for XML 1.1 without the appendixes.
Now consider that SGML gives poor-man's versions of schemas,
namespace (SUBDOC), regular fragmentations (shortrefs), and
compression (minimization), and a profile/feature system to allow
the user to view things as layers. It is nice to remove 33 pages
from SGML: but we have replaced them with hundreds of pages
of specs like namespaces and XML Schemas. Viewed as an exercise
in reducing the impenetrable, XML has been a total failure; viewed as
an exercise in refactoring, it has not been a total failure.