Re: Browsers and syntax errors (Was: Captions for Figures...)

Steven J. DeRose (sjd@ebt.com)
Fri, 10 Mar 1995 18:57:16 +0500


> While I'm here - can anyone pls tell me, why *do* I need to keep
> turning <b> on and off, to skip over paragraph, list, and other
> such tags? Some documents, like some of the WDVL pages, I really
> *do* want to be bold. Presentation issue, I know, and hurrah for

Do you mean you want a whole subtree to be bold? If so that's still a tree
and SGML has no problem with it (although HTML chose to rule it out). I
consider this HTML decision wise. If there's a *reason* you want something
bold, then the document's markup should express the reason, not just
"bold".

If you force the whole document to be bold, you're not expressing the
presence of a meaningful object anyway, but just imposing on the user. You
wouldn't force them into a particular font and size or into all-caps,
right? Those may be tempting for some occasions, but would hose people who
don't own that font, or can't read without larger type, or who stuck on a
monofont device. So why permit forcing 'boldness'?

If you instead meant that you want some random span to be bold, like the
last part of one paragraph and the first part of the next, that's a
fundamental issue because the SGML/HTML document model is that of a tree,
not an unstructured steam with embedded formatting commands like troff.
Trees allow you to process documents in much smarter ways (like, you can
re-render without backtracking to the beginning every time there's a little
change, you can make hierarchical navigation tools like outliners, you can
do far smarter information retrieval, etc.)

Asking why <b> can't span across paragraph boundaries would be like asking
why this expression is no good in basic algrebra:

( a + [ b * c ) - 4 ]

Steve