Re[2]: proposed new Media Type application/iges

Alan Peltzman (peltzmaa@CC.IMS.DISA.MIL)
Mon, 20 Mar 95 17:24:40 EST


3D Colleagues:

Can we initiate an effort to introduce a
new MIME type that can accomodate
3D vector graphics?

I would propose a more general category than just 3D:
graphics/product data

/or/

graphics/product data modeling

In any event, I believe that CGM and IGES need to occupy
the same MIME type.

- Alan Peltzman
______________________________ Reply Separator _________________________________
Subject: Re: proposed new Media Type application/iges
Author: linas@innerdoor.austin.ibm.com (Linas Vepstas) at smtp
Date: 03/20/95 04:29 PM

Hi,

> > We are working toward proposing "world" as part of the VRML effort.
> > VRML is a 3D extension to the Web. Check out http://vrml.wired.com/
> > The prototypes are proceding with x-world/x-vrml (and the less polite
> > using simply world/vrml)
> >
> > I strongly believe that the 3D model is a fundemental datatype. IGES is
> > one, DXF, CGM (with 3D extensions), dof, off, iff, iv as well as
> > application specific files.
>
> 1) do 3D models require different "hardware" support or support from
> the transmission medium to be usable, from (say) that for image/* or
> application/*?

Yes. Absolutely. The point is that a old 286 can display gifs and
other pixmap formats at high speed, and quit well. Not true for 3D.
Even pentiums & power-pcs are sluggish handling 3D without h/w acceleration.

Someone used the words "mime" and "pager" in the same breath. In that
case, we should mention "world/*" and "game machine" or "set-top box" in the
same sentance.

The point here is that a sharp tongue could have argued that post-script is
a kind of "image", or that an mpeg movie was a kind of "image". They're not,
and neither is vrml. In almost every fundamental way, its different. Don't
loose the forest for the trees.

> 2) is there a default user agent behavior that applies to *all* 3D
> models in your proposed world/* class?

(Potentially) yes, although this is funky. We are talking about adding
interaction to vrml v.2.0. Implementing that interaction almost certainly
means opening up a new/different communications path between browser
& daemon (and/or cgi). Potentially, all interactive media would want to
use this scheme. Whether this is an argument for a new, fundamental mime
type, I dunno.

> The top level in MIME types is designed to allow filtering by gateways
> between different kinds of messaging systems (so that you don't try to
> ship audio/* to a pager or image/* to voice mail), and also to provide
> default user agent handling of content-types. So the argument needs
> to be pitched in these terms.
>
> Keith

--linas