Re: Supposed VRML files

[email protected]
Thu, 25 May 95 21:02:41 CDT


> >
> > Dog List: http://www.geom.umn.edu/~daeron/docs/dog.html
> > Legit List: http://www.geom.umn.edu/~daeron/bin/legitlist.cgi
> > Authenticator: http://www.geom.umn.edu/~daeron/docs/vrml.html
>
> PLEASE don't hang our dirty laundry out for all the world to see. Putting
> an innocent author up for ridicule by including his or her home page into
> your personal black list is mean spirited and doesn't help ANYONE. It
> does, in fact, tell the world that we are divisive group who can't behave
> ourselves.
>
> IvToVRML has a problem, it leaves the nearDistance and farDistance fields
> in a camera, as well as a couple of other things. This is a bug which
> will be fixed in the next beta, not a blatant attempt by SGI to take over
> VRML! I have seen several authors change the header on an IV file to VRML
> and call that VRML. Making these authors into criminals will only serve
> to scare off those who really want to jump start VRML.
>
> Your authenticator is GREAT! Please promote that. But if you encounter
> an illegitimate file couldn't you e-mail the author and suggest he or she
> fix the problem rather than holding him or her up for derision?
>
> WE MUST PLAY NICE OR THE WORLD WILL IGNORE US...
>

<long sig CHOMPED>

Hello Chris,

Thanks for your input. You make some very good points in your letter,
and I have changed my practices accordingly. From now on I will send
an email message to each information provider before I add their URLs
to the dog list and give them a day to fix it. This brings me to my
next point. You have an object linked from your home page that is
not legitimate VRML, even though the link is titled "VRML". The
authenticator informs me that:

For URL: http://reality.sgi.com/employees/cmarrin/Museum.wrl

Parsing with QvLib revealed:

VRML read error: Unknown field "nearDistance"
Occurred at line 15
This is *NOT* a legitimate VRML file:

Please fix it or risk being added to the dog list. :-) hehe...
Perhaps, you need to take things less seriously though. In defense
of the dog list, I'd like to say that it is certainly not a black
list, but rather a means of encouraging folks to be less lazy.
Maybe you don't know how frustrating it is to spend many days working
on a VRML browser just to find, when your almost done, that all the
most interesting objects out there aren't VRML even when they claim
to be so.

This is a completely invisible problem to WebSpace users, but to
those of us building alternative browsers, it is a critical issue.
Hence it is important to raise awareness about the problem of
nonstandard VRML. I'm aware of the problem with the inventor to
VRML converter and I'm glad to hear it's being fixed but why not
patch up your files in the meantime with something like:

cat bogus.wrl | grep -v nearDistance | grep -v farDistance > better.wrl

Then you can comment out all the other non-VRML things as well if
you aren't afraid of using a text editor. :-) What really appalled me
was some discussion I saw earlier that read something like this:

ButtHead: uh... I want to make a VRML object... that would be cool...
huhuh huhuh...

Beavis: hehe hehe... yeah yeah VRML rocks!! Hehe... I think you can
do that by, like, replacing the header in an Inventor
file with "#VRML V1.0 ascii" hehe hehe...

ButtHead: cool!... huhuh huhuh, were there dude!

What's really annoying, is that WebSpace actually lets you get away
with doing just that! And no, I don't think SGI is trying to take
over VRML, I just think it was poor judgment to release a VRML viewer
that plays so fast and loose with the language. I like SGI; I use their
hardware and software every day at work. They should just be more careful
in their position of influence. I'm looking forward to the day when
SGI and others clean up their act and serve pure VRML. Then there won't
be a need for the dog list (http://www.geom.umn.edu/~daeron/dog.html).

Peace,
Daeron

(this letter written completely by voice)