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

Brian Behlendorf ([email protected])
Thu, 15 Jun 1995 12:11:16 -0700 (PDT)

On Thu, 15 Jun 1995, Andrew C. Esh wrote:
> I think you guys are missing the point. I may be misinterpreting the
> original suggestion, but I did not get that impression that it was
> intended as a requirement, or that any browser should use the site name to
> decide how to display the received files. The point of the suggestion, as
> I understood it, was to help general users identify, by host name, which
> was the best server to be trying to get files from. For example, if I were
> to get a hint from someone that Netscape Inc. has a cool new VRML space,
> then my gui> at where to find it would be "vrml.netscape.com", just as my
> gui> es would be tend to be correct for FTP and WWW from the same site.

I fully understand it was just a suggestion for a convention, not a
standard, I just thought that it wasn't as elegant as it could be, since
it semantically puts "WWW" and "VRML" as mutually exclusive, whereas the
latter is just a part of the former. It would also cause those who want<
to be "serious" about VRML to fire up a separate server on a separate IP
number (virtual or real).

I suggest that instead of http://vrml.host.com/ the convention be

http://www.host.com/vrml/

Whether that's really where the vrml files actually live or not doesn't
really matter - the server could just provide an index page (or "museum")
of vrml objects on the server wherever they might be. Most sites won't be
purely VRML, they will be an integration of VRML and HTML and GIF's and
PDF's and Java apps, so it would make ense to keep links relative and
have all those media on the same machine. The suggestion above is also
consistant with a couple other conventions - /classes/ in Java, /Admin/
in NetSite administration, etc.

Brian

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