Re: my two more 2 cents on VRML (was Scripts vs. APIs)

Brian Behlendorf (
Fri, 9 Sep 1994 16:49:35 -0700 (PDT)

Linas (and others who have joined the vrml list recently...)

I have been delinquent in my duties as list moderator - as happens with most
useful net-community-based projects it seems, I have only been able to do
this in my spare time, which had become exceedingly rare over the past few
months while I've been coding up a storm here at Wired. I call myself a
moderator, yet I don't approve posts before they are sent - I prefer the free
flow of ideas. However, as *any* who has spent time on the net knows, that
form of unstructured discussion can quickly lead to fragmentation unless well
monitored. Basically I revisited my vrml folder to find a raging debate
going. Either Mark or I or someone should have stepped in and added some
context, so here is some.

Many of the issues that are being discussed (APIs/scripts, network
caching, etc.) have been hashed out long ago. 6 weeks ago I put up a
survey to guage the opinion matrix of the list members - what features
should definitely be in the first cut, which ones should be delayed, and
which ones we probably should never worry about. Three weeks ago I
compiled and posted the results of that survey. It's available at:

I have yet to truely convert it to proper HTML, but Hypermail converted
the important elements correctly.

I *strongly* suggest that anyone who was unaware of the conclusions of
that analysis look at that document, and the other archives for that

One result of the survey was that Open Inventor, Autodesk's CDF, and OOGL
were identified as the most attractive description languages that might
fit into the mold. It was suggested that those groups put up as much
info about themselves as possible (OOGL has a fantastic web setup, and
Inventor's is great too) and that the other members look closer at this
narrowed field of contenders, and that at some point in the near future
one of those would be chosen as the VRML basis. At that point the list
would work to refine the chosen language to incorporate everything
identified as necessary for the first cut of VRML.

Instead, discussion has diverged into more "should" statements. I take
blame for that - workloads here had caused me to simply be very liberal
with the "d" key.

I haven't even looked at the most recent flare-up, though I'll post on
that too.