Re: SPEC: DNS conventions & "naming" of VRML sites

Brian Behlendorf ([email protected])
Wed, 14 Jun 1995 13:53:43 -0700 (PDT)


On Wed, 14 Jun 1995, Chris Marrin wrote:
> On Jun 14, 11:28am, Brian Behlendorf wrote:
> > Subject: Re: SPEC: DNS conventions & "naming" of VRML sites
> >
> > Actually, it's very easy to return a VRML document to a VRML browser and
> an
> > HTML document to an HTML browser for the *same* URL (i.e.,
> > http://www.vrml.org/), given that browser authors are doing the right
> thing
> > with Accept: headers. (HINT HINT) I also think "WWW" includes VRML and
> all
> > other media types, so I don't agree that we need a separate convention
> for
> > naming HTTP hosts that hold VRML pages.
>
> Not being guilt ridden I won't assume that HINT referred to any browser in
> particular :-)...

I haven't tested any of the VRML-particular browsers on this, I would
presume you were more aware of this than anyone. I was mostly referring
to the other regular web browsers. Emacs-W3 and Arena are the only ones
that come close.

> A while back I tried using "Accept: x-world/x-vrml" and had the http
> server I was connecting to send me html anyway!

Depends on what representations existed of the object you were requesting
(as well as what server was being used). If (for example in one
way to configure it for Apache) there's an index.wrl and an index.html,
then with a request for "index" should return the .wrl file contents if
you set the Accept: line as above. I wish I had an example to try out
for you - we will soon.

> I talked to one of our
> experts and he laughed, saying "yeah, nobody supports that".

CERN has supported it since the early days, and Apache has it now too.

Brian

--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
[email protected] [email protected] http://www.[hyperreal,organic].com/