Re: [x3d-public] NullPointerException when loading scene with script node

[Braden McDaniel <braden@endoframe.com>]

[Sorry for the delay in responding; I had to recover from a botched
Fedora Core 5 upgrade.]

On Wed, 2006-03-22 at 14:23 -0800, Alan Hudson wrote: 
> Braden McDaniel wrote:
> > Alan Hudson wrote:
> > 
> >> Braden McDaniel wrote:
> >>
> >>> Alan Hudson wrote:
> >>>
> >>>> George Anastassakis wrote:
> >>>>
> >>>>> Hello all
> >>>>>
> >>>>> I'm trying to load a scene with a Shape and a Script node but I 
> >>>>> cannot get it right. The file is:
> >>>>>
> >>>>> #X3D V3.0 utf8
> >>>>> PROFILE Immersive
> >>>>> DEF s Shape {
> >>>>> appearance Appearance { }
> >>>>> geometry Box { }
> >>>>> }
> >>>>> DEF sc Script {
> >>>>> eventIn SFTime test
> >>>>
> >>>>
> >>>> this is not valid X3D.  it must be inputOnly
> >>>
> >>>
> >>>
> >>> We've discussed this already and it was made clear that this *is* valid
> >>> X3D. A conforming processor will not reject it. "eventIn" is a 
> >>> (deprecated)
> >>> synonym for "inputOnly".
> >>>
> >>
> >> That is your opinion and not consortium policy.
> > 
> > 
> > It's not my opinion. It's what the specification says. "Consortium policy"
> > (whatever that is) is irrelevant. An ISO specification is at issue here;
> > its interpretation is not a function of "consortium policy".
> > 
> The consortium is the place where edits to that ISO document are 
> generated.  If there are issues of interpretation then the consortium 
> helps disambiguate them.

And if there were ambiguous language--i.e., language that could
reasonably be interpreted more than one way--that would be appropriate.

But that's not the case here.

The specification clearly says that "eventIn", "eventOut",
"exposedField", and "field" are "deprecated keywords".

> >>  I agree the language needs to cleared up in that area.  But authors 
> >> who want their content to work in all X3D browsers will use the 
> >> correct keywords.
> > 
> > 
> > Please stop using such pejorative (and incorrect) language. The deprecated
> > keywords are no less "correct" than the new ones; they are simply
> > deprecated.
> > 
> These keywords are not in the grammar at all.  Hence there is an 
> ambiguity in what the ISO spec means.

Calling this an ambiguity is not an accurate characterization. The
absence of these keywords in the grammar is an *inconsistency*, which is
distinct from an ambiguity. Whereas an ambiguity is a case where the
same language can reasonably have more than one meaning, an
inconsistency is where different portions of language in the spec has
incompatible meanings.

In this case, the grammar is inconsistent with a lot, not the least of
which is itself. It has a sufficient number of errors that a reader can
quickly infer that it is not reliable.

> http://www.web3d.org/x3d/specifications/ISO-IEC-19776-X3DEncodings-XML-ClassicVRML/
> 
> 
> Tools and content authors should not use the keywords.

That's inaccurate. Per the language in the spec, conforming tools must
support these keywords. That's what "deprecated" means. "Deprecated"
does also mean that content authors should indeed be discouraged from
using them; but such discouragement need not include mischaracterization
of the keywords as "incorrect".

>   In the future 
> they will be removed from the spec completely.  I expect that will 
> happen on the next revision of the standard this year.

Given that the last time this came up there was not apparent consensus
on the issue among the spec authors, I'm not inclined to put a lot of
stock into this prognostication. Certainly if consensus *has* been
achieved, it would be very helpful if a summary of intent were posted.

That said, certainly tool vendors would like to retain compatibility
with the current specification if the Consortium opts to break
compatibility in a future revision to the specification.

>   So encouraging 
> users to use those keywords is a bad idea.

As far as new content is concerned, I agree. And so is propagating
misinformation about the keywords' status.

-- 
Braden McDaniel <braden@endoframe.com>

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