[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [www-vrml] Why you _can_ convert VRML97 to X3D
Goncalo N M de Carvalho wrote:
(I'm not getting all the emails for this thread. Only
commenting on what is written below)
Any (automatic) conversion of a script node that
relies on dynamic generation of content in classic
VRML encoding is non trivial.
If this type of script behavior is only 1% of all
content I don't know. Personally I don't think anyone
does.
my estimate was quantitave. 1015 javascript: VRML97 scripts
converted to ecmascript: X3D scripts in the content archive.
Only 10 or 20 scenes needed manual intervention to work OK.
Scripts that load content (VRML) need to provide the
conversion themselves for full portability.
I believe the construction calls are exactly the same,
tho haven't tested it yet. Example available at
http://www.web3d.org/x3d/content/examples/course
CreateVRML From String Random Boxes
The additional existence of the PROFILE statement and
the fact that is being checked at parsing time also
doesn't help. Assuming that the browsers do enforce
it. Some popular ones don't.
Checking at parsing time also allows for silly
examples
like this
#X3D V3.0 utf8
>
PROFILE Full
Shape { }
wondering what is wrong with that. sometimes trivial
files like this are used as Inline or run-time loader
fodder. so allowing trivial examples has some value.
it's good that allowing a higher profile is forgiving
because it eliminates the need to frequently recompute
the profile. sometimes a scene will load higher-profile
content at runtime into that scene, it's good to have a
higher ceiling.
It would be more useful to enforce it after parsing.
(AFAICT the spec doesn't make any requirement on when
the enforcement (if any) should be made.)
From the spec and section 7.2.5.3 it is unclear that
that the browser needs to respect it at all:
"...and providing a mechanism for strict conformance
should the browser choose to enforce it..." but then
it says
"it may proceed with presenting the world; otherwise,
it shall fail."
browser conformance is always good. some day, maybe even
certified.
BTW there are examples in the spec that do not use the
PROFILE statement at all and others that define it
like this PROFILE "Full". Which one is it?
possibly a problem here - not sure if we have defined
a default profile. that might be an erratum in Amendment 1.
There is the question of portability that Paul
mentioned. That is a most serious one.
If I have a company with *lots* of VRML content I
won't pay to convert it. Why should I? The whole
premise of ISO is backwards compatibility. If the new
stuff doesn't support it I'll use the old stuff for
it.
your point is well taken. still, many folks will want to
upgrade it, particularly for conversion to .x3d encoding so
that XML benefits can be realized.
What is the purpose of X3D-Edit's warnings? If it is
to enable a author to go and alter the content by hand
then its pointless.
What's done with the warnings is up to the author.
We go back by hand and fix them. They are also used
during regression testing to ensure that new errors
haven't been induced or undetected in past content.
> Decent 3D content needs to be
built in a DCC tool or with a suite of tools.
The whole VRML build-by-hand world is dead.
We get a lot of progress using X3D-Edit. We're glad
that VizX3D and other tools are out there.
XML-aware editing is particularly useful for
trying out new patterns and checking autogenerated
X3D content.
Perhaps our content is indecent or dead, but it
doesn't feel that way!
Besides, command-line tools are not good enough.
Maybe I'll start a simple converter of my own.
go for it. or work with some of the emerging ones
and improve them.
thanks for the great insights and dialog Goncalo.
all the best, Don
--
Don Brutzman Naval Postgraduate School, Code USW/Br work +1.831.656.2149
MOVES Institute, Monterey CA 93943-5000 USA fax +1.831.656.7599
Virtual worlds/underwater robots/X3D/XMSF http://web.nps.navy.mil/~brutzman
-------------------------------------------------------------------------
for list subscription/unscrubscription,
go to http://www.web3d.org/cgi-bin/public_list_signup/lwgate/listsavail.html