Default entities and AT&T

Jonsm@aol.com
Mon, 24 Jul 95 02:13:17 EDT

After browsing around the Web with a SGML based syntax checking browser I've
come to the conclusion that there are 1,000s of cases of inadvertent entity
use in phrases like AT&T and embedded URLs. The proper way to fix this of
course is change all of the &'s to &amp but this isn't likely to happen. So
instead I looked into changing the parser.

This is when I discovered the concept of default entity:
<!ENTITY #DEFAULT "Warning: The entity is not defined!">
If this declaration is not present then the default default entity is null.
Upon parsing AT&T turns into 'ATWarning: The entity is not defined!' or just
'AT'.

The question is: should I change the behavior of the default entity and make
it return &T for the unknown entity? If so, the HTML spec should note this
behavior is a deviation from standard SGML. Conversely, a browser that
returns &T would be non-conformat to the HTML2 spec since HTML is a SGML
application.

I read through the sections in the HTML2 spec on Data Characters and markup
errors and I didn't see any reference to the problem.

Jon Smirl, jonsm@aol.com