Actually, apart from the fact that some of the extensions were contrary to
the SGML model that people have been spending time squeezing HTML into, the
problems many people have with NetScape's extensions is that they are billed
as "things not in the official specs but they soon will be"... by encouraging
people to use them as standard tags and suggesting people call their pages
"Netscape Enhanced" they are attempting to direct market pressure on the HTML
spec process. If they had been labeled "experimental" and told people "use
at your own risk" then it would have been less of a problem. People *need*
to try out new tags before defining them as a standard of course, so this
sort of experimentation should be encouraged - but when experimental
practices become the de-facto standard, that's a recipe for danger. It's
especially worse in HTML, where the Netscape browser is clearly the
predominant one, and where many people learn HTML by looking at other HTML
pages they admire. I wonder what will happen when NetScape *isn't* the most
> For example, part of the Inventor group at Silicon Graphics is busy
> working on VRML-related stuff. What problems will be caused if we
> produce a VRML browser capable of reading all Inventor nodes, and not
> just the VRML draft spec nodes?
> -- My thoughts: The problem is that authors might unknowingly produce
> VRML worlds that could only be read into Inventor-based browsers.
> That would be bad. Would a "VRMLLint" program that checked a VRML
> file for correctness and use of only official VRML nodes, plus an
> Inventor to VRML translator program be enough to alleviate the
In my opinion, yes.
The existance of standards shouldn't stifle the drive to improve things -
it simply serves as a solid base so that a browser author can say "my
browser is fully VRML-1.0 compliant" and a VRML author can say "this page
is 100% VRML complaint", and the author can rest assured that his VRML
page will be presented just as he wanted on all VRML 1.0 browsers. If
the browser can interpret other things above and beyond the standard set
of VRML, that's fine.
firstname.lastname@example.org email@example.com http://www.hotwired.com/Staff/brian/