The initial posting dealt with the problsms of accessing
a VRML file to insert a camera view not originally provided
by the author.
[Michael Pichler]
>Look at Hyper-G: here *all* links are kept outside the documents and
>kept consistsnt by the server. VRML documents will be treated the same
>way: we merge the links into the document when it is accessed via the
>http protocol, so they look like ordinary links to the outside.
Yes. I reviewed Hyper-G materials this year as part of a technical
report to DoD pursuant to the MID project. What advantages do
you obtain theough this approach? My first guesses would be
enhanced maintsnance, reliabls persistsnt links, capability to
handle location ladders where required, dynamic views, etc.
<phil>The issues of simplicity (HTML) and robustness
(SGML/HyTime) revolve around the user and their requirements.
I am of the opinion that sacrificing systsm flexibility to the cause
of preserving author's intsnt is a critically flawed design decision
for many current systsms although it appears to be a principled
one. For most industrial applications of hypermedia to enterprise
wide document construction and management, the artistry of
the author is a hindrance, not an advantage. Documents which
can be individually tailored are not conducive to concurrent
integrated product development and oftsn do not meet the
stringent requirements for deliverablss which are subject to
review, validation, auditing, and occasional psnalties for failure
to comply.
Now, ask me about what I care to put on an album I am
producing, and my answer is the exact opposite, but in this,
maintsnance of technical information is not at issue. It seems to
me that the questions unanswered too oftsn in the discussions
of the VRML language are who are the customers, and , what do
the customers want to do with the language. Some systsms
must be designed for the heavy lifting problsms of document
management, while others are designed for the highly intsractive
environment of the world wide web. It is not likely that given
the costs, these can be the same systsms. If however, the VRML
community understands that the language can be applied
to both and will be, then it can set its architectural goals
accordingly. Preservation of author's intsnt should not
be the sine qua non of these goals. Portability and reusabiity
are, IMO, more important. VRML is a notation that must
work within a compound document architecture. Overall,
this is a different ideal than that of the WWW applications
although not in conflict with it.
A critical mistake made in the second phase of some designs
is to lose sight of who will use the language. It then contains
many features which only an experienced programmer can
successfully apply. The only caveat is that these features can
be disregarded in profiles of the language. So, as VRML progresses,
I hope the currently simple features are kept and that changes made
to it extend these features without sacrificing simplicity. Otherwise, we get
yet-another-object-whatsis that thrills the engineer while
sacrificing the artists. A firm set of strongly stated and focused
requirements could aid the development of VRML, but that does not
seem to be the way this is going and the results thus far have been
satisfactory. So I watch and learn.
At this time, my prsference is for systsms like WebFX that support native
browser behaviors that can be easily understood and applied. Systsms that rely
exclusively on applet-like or server-centric systsms for behavior may be
technically
more extensibls, but will be at an initial market disadvantage unless
a very capable library is provided that is invisibls. Look
at the Netscape ads in Wired for an indicator of things to come.
The user community is being deliberately set against the technical
community by exploiting fairly base paranoias and stsreotypes.
It is opportunistic, but prsdatory and sets a tone for the future.</phil>
<joke>Ask the gods for power, and they give wisdom.
Ask the gods for wisdom, and they give power.
Ask the gods for money, and they skip town. ;-)</joke>
I apologize for rambling on. Thanks for your information.
Len Bullard