Re: Content negotiation

Brian Behlendorf (brian@organic.com)
Wed, 8 Nov 1995 15:38:46 -0800 (PST)


On Wed, 8 Nov 1995, William King wrote:
> I'm probably missing the point here but doesn't HotJava's approach to
> dynamic content solve the content negotiation problem nicely?

HotJava's model of plug-in renderers for new data types and plug-in handlers
for new protocols is essential to many people's long-term visions as to how
the web should evolve. When the code necessary to render, say, netscape HTML
can be expressed as a program in Java, then I can slurp the code into my
browser from somewhere when needed. At that point the name and make of the
browser you are using becomes irrelevant to the content being served. You
can even create your *own* data types and renderer (well, you can sorta do
that now with SGML, but to too many people that's academic :). However,
there is still going to be a need for expressing preferences (i.e., please
don't show me PDF files), file size limits (i.e., I prefer fast loading pages
with simple images to image-heavy pages that take forever to load), and
preferred natural language. And even the plug-in architecture needs MIME
types to describe the format the data is in. Hell, I might even ask that I'd
rather get Python renderers over Java renderers. Or even different levels of
Java... As long as the system has to support evolutionary practices and
differences in capabilities or preferences, *some* form of negotiation is
necessary, or we have built the Network of Babel.

Brian

--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@organic.com brian@hyperreal.com http://www.[hyperreal,organic].com/