VRML help needed

Hi Guys -

I started a very simple VRML importer but I haven't gotten very far yet
and was wondering if there is any interest in helping out. The current
code is in the contrib repository at

   http://hedgehog.software.umn.edu:8888/Contributions.html

So if anyone's interested grab the current version from there and help
improving it!

Cheers,
   - Andreas

Hi David -

Since I've never looked at your code I can't really comment on it but
all I did was a very simple adaptation of the old (B3D) based importer.

Cheers,
   - Andreas

David Faught wrote:
> I was thinking that it wouldn't really take too much to redo the one that was working for Jasmine, other than coping over some of the classes from the old Wonderland stuff and modifying the references to use the new matrix classes.  Have you started all over with a different approach for some reason?
>
>

Sure. Two reasons mainly: #1: I have never heard the name Collada before
you mentioned it. And even right now (since I haven't googled it yet) I
don't know what it is and how it relates to VRML or Croquet. Care to
elaborate? #2: I reused 99% of existing code. I spent exactly two hours
on the VRML importer mostly to get a feel for how hard a forward port to
Hedgehog would be. Turns out it wouldn't be hard at all.

Cheer,
   - Andreas

Joshua Gargus wrote:

Since that is basically what I did for the Jasmine importer, it sounds
like you are a step ahead of me for a Hedgehog version.  The Jasmine
importer is still on SqueakMap with a description page on the old
Croquet Wiki.

On 4/24/06, Andreas Raab <[hidden email]> wrote:

Hi David -

It'd be great if you could go over it and "compare notes". Even more
valuable would be to find a collection of VRML files that we could use
to test various features and figure out which of the (many) things are
still missing.

Cheers,
   - Andreas

David Faught wrote:

There are still (so far) some of my notes about what didn't work on
the old Croquet Wiki at
http://minnow.cc.gatech.edu/croquet/VRMLImporter

There are also some test files there, but ones that mostly worked.

I haven't looked in detail yet, but the things that I remember fixing are:

1. The cache for primitive objects on the class side of the old B3D
importer didn't work properly because it gave back a reference to the
actual cache object rather than a new copy of it.  Although I took the
slash-and-burn approach to fix this, a well placed "copy" message
would probably work better.

2.  The primitive object methods didn't conform to the VRML standard
for the sizes of the untransformed objects they returned.  I don't
remember the specifics, but a sphere should have a radius of 1 unit,
for example.

3.  Orion complained that sometimes his textures would come out in the
wrong orientation on the imported objects.  Again, I used a brute
force approach and forced a recomputation of the texture coordinates,
ignoring what was in the file.  Not good, but it worked for most
things.

4.  I vaguely remember fixing something with the texture file
reference itself and where it was in the directory structure relative
to the VRML file.

5.  Orion provided a patch that fixed some hard-coded limitation of
500 somethings.  I don't remember the details of this one either, but
the patch is still in the old Mantis entries somewhere.

I'll check these out as I have time.

On 4/24/06, Andreas Raab <[hidden email]> wrote:
> Hi David -
>
> It'd be great if you could go over it and "compare notes". Even more
> valuable would be to find a collection of VRML files that we could use
> to test various features and figure out which of the (many) things are
> still missing.
>
> Cheers,
>   - Andreas