Re: Vendors taking SGML seriously

Amanda Walker (amanda@intercon.com)
Wed, 12 Apr 95 22:35:11 EDT

> This is a fascinating statistic, you are saying that the effort of supporting
> illegal HTML has tripled the cost of your development effort...

Yup, that's about right. We did manage to run a very fast development cycle
on our browser (4 weeks of design, 2 weeks to full implementation of
Level 2 HTML 2.0 (according to the draft as of 11/94), 4 weeks cleanup on
real-world pages). However, the fact that our rollout was at MacWorld Expo
in January instead of Email World in December was principally a result of
having to fix up our behavior when presented with bad input, by having our
QA people surf the web looking for "ugly" pages, and then comparing our
rendering with other browsers to determine what our workarounds should be.

> how much of
> this was SGML-compliance issues, i.e. tag form, character set, etc. ? The
> answer should provide sufficient motivation for small guys to stick with SGML.

Most of it has been basic syntax and embdedding issues, all of which can be
easily attributed to authors composing HTML by hand but without an actual
understanding of the basic ideas behind SGML (or structural markup in
general). For example:

- Omission of the trailing '"' character in quoted strings.
- Terminating symbolic character references ("&XXXX;") with
whitespace instead of ";"
- <DT> with a <DL> inside of it
- <DT> without a corresponding <DD>
- <DD> used outside of a <DL> (ostensibly to provide "indented
paragraphs")
- <A> emdbedded within another <A>
- Bogus URLs (my favorite is "d:\www\foo.html")

and so on.

Amanda Walker
InterCon Systems Corporation