Re: VRML 2.0 v old wrl loader

New Message Reply Date view Thread view Subject view Author view

Angus Dorbie (dorbie++at++sgi.com)
Thu, 14 Oct 1999 12:57:26 -0700

Performer has always been loader agnostic and the whole concept of file
format loading in the API is that the loader reads a file and creates a
Performer scene graph. The scene graph is the defining format, the
loader itself is often supplied by 3rd parties.

The de facto 3D format for Performer if there is one would be Open
Flight, that works very well and is provided by a 3rd party. In that
instance the relationship works extremely well and the loader is
functional and excellent. So much so that other 3rd parties dropped
their own in house Open Flight loaders. In addition other tools vendors
adopted Open Flight as a major format.

The VRML loader could work better but you have to cooperate with the
vendor. The other observation you make and it may well be valid is that
VRML 2.0 is overly complex and not well suited to the applications and
markets Performer serves best.

Cheers,Angus.

Anton Koning wrote:
>
> Hi Performers,
>
> As some or perhaps most of you know the old .wrl loader is a VRML 1.0 (more or
> less identical to the Inventor format), while the new .wrl loader claims to be
> a VRML 2.0 loader (which is complete different from Inventor). Fortunately you
> can still use the Inventor loader to load most VRML 1.0 files by changing the
> extension to .iv.
>
> We at SARA use Performer for some of our CAVE applications and have had major
> problems with the OpenWorlds loader. It chokes on the most simple VRML 2.0
> files and while we have reported some problems (some of which have been fixed)
> we still find the loader unusable. For example the fact that certain versions
> of the loader only function when used with Perfly is a sign of inferior design
> and little or no testing.
>
> The fact that SGI, who is one of the major players in establishing the VRML
> 2.0 standard, relies on a third party to provide a Performer loader is in my
> opinion a big mistake. The fact that the loader as provided is unusable and
> apparently completely untested is inexcusable.
>
> In defense of OpenWorlds I have to admit that is seems almost impossible to
> write a complete and robust VRML 2.0 loader. Although many packages write VRML
> 2.0 code, few implement a loader and even fewer have a good (let alone
> perfect) loader. All this leads us, after 2.5 years of struggling with various
> VRML 2.0 datasets, to the conclusion that VRML 2.0 is (still) unusable as a
> standard for 3D polygon data.
>
> ------------------------------------------------------------------------------
> | Anton H.J. Koning, Ph.D. | Consultant, Academic Computing Services Amsterdam |
> | e-mail: anton++at++sara.nl | s-mail: P.O. Box 94613, NL-1090 GP Amsterdam |
> ------------------------------------------------------------------------------
> -----------------------------------------------------------------------
> List Archives, FAQ, FTP: http://www.sgi.com/software/performer/
> Submissions: info-performer++at++sgi.com
> Admin. requests: info-performer-request++at++sgi.com

-- 
"Success is the ability to go from one failure to another with
 no loss of enthusiasm."    -    Winston Churchill.


Performer + OpenGL examples and tutors: http://www.dorbie.com/

New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Thu Oct 14 1999 - 12:57:32 PDT

This message has been cleansed for anti-spam protection. Replace '++at++' in any mail addresses with the '@' symbol.