Re: [www-vrml] Why you shouldn't convert VRML97 to X3D

["George Birbilis" <birbilis@kagi.com>]

The idea is though that for one to write such utility they have to do lots
of research first on both VRML and X3D. If a step-by-step manual convertion
guide was available from Web3D, people could then write such utilities
easier

> In any case, this is a job for a utility if you're dealing with complex
> files. The simple cookbook Leonard recommends can get a person started
with
> their old VRML files and just hand-editing. But it's not a reliable
> production solution and doesn't cover the hard corners that a serious
> application might hit.
>
> In addition, I strongly disagree with the following
>
> > The bottom line is: do not convert from VRML97 to X3D.
>
> Have you read that other thread? There's like a million VRML files out
> there. Assuming that only 10% are good (:-o) that's still a lot of content
> that people don't want to leave behind when moving to X3D. And there's
> really no good reason to. As discussed, most of the conversion can be
> automated. And what can't is program code-- which often breaks when you
> convert up anyway.
>
> So, we're not in bad shape here.
>
> Tony
>
> > -----Original Message-----
> > From: owner-www-vrml@web3d.org [mailto:owner-www-vrml@web3d.org] On
Behalf
> > Of c3099023@tees.ac.uk
> > Sent: Sunday, September 05, 2004 6:02 AM
> > To: www-vrml@web3d.org
> > Subject: [www-vrml] Why you shouldn't convert VRML97 to X3D
> >
> > Consider the following
> >
> > #VRML V2.0 utf8
> >
> > Inline
> > {
> >    url "some.wrl"
> > }
> >
> > Assume that "some.wrl" points to a well-formed portable world. A fully
> > compliant VRML97 browser will render it correctly.
> >
> > Use the official recipe for conversion that is being advertised
> > and you get
> >
> > #X3D V3.0 utf8
> >
> > Inline
> > {
> >    url "some.wrl"
> > }
> >
> > You try to run it on a X3D browser fully compliant with the Full profile
> > and you get nothing. Why? Section 9.4.2 of the spec says this world may
> > produce undefined behaviour.
> >
> > Another example:
> >
> > ...
> > Switch
> > {
> >     choice [ ... ]
> > }
> >
> > Convert it with the official recipe and you get a syntax error. Choice
is
> > not a field of Switch in X3D.
> >
> > Note that I'm not even touching Script nodes. That's just too much fun
for
> > a single email. SMTP won't handle it.
> >
> > The official recipe is nothing but a nasty, ugly, non-portable hack.
> >
> > The bottom line is: do not convert from VRML97 to X3D. If you must make
> > sure you're not falling in dark corners of X3D spec. X3D is *not*
> > backwards
> > compatible with VRML. Only use X3D for new content. If you rely on a lot
> > of scripting for which you only have bytecode, then stick with VRML.
> > Nothing wrong with that.
> >
> > Cheers,
> >
> > Goncalo
> >
>
> -------------------------------------------------------------------------
> > for list subscription/unscrubscription,
> > go to http://www.web3d.org/cgi-
> > bin/public_list_signup/lwgate/listsavail.html
>
>
> -------------------------------------------------------------------------
> for list subscription/unscrubscription,
> go to
http://www.web3d.org/cgi-bin/public_list_signup/lwgate/listsavail.html
>
>

-------------------------------------------------------------------------
for list subscription/unscrubscription,
go to http://www.web3d.org/cgi-bin/public_list_signup/lwgate/listsavail.html