>> The OpenGL process and OpenES process
both seem to incorporate more contributors then X3D with good results.
True - every group I have seen in Khronos has an natural
onion ring participation model.. OpenGL ES has about 50 companies
lurking on the Working Group list, about 20 of those turn up to meetings,
10 of those actively contribute to the debate - and about 5 companies get
involved in editing and creating documents and code.
I think you can extract two principles here: a)
strong industry participation does NOT slow down the decision process -
and it DOES help momentum when those 50 companies execute fast-track product
launches based on having early access to the specs as they are being created;
b) even if all 50 companies are not inputting - the fact that there
is an open process - and all information and decisions are transparent
to all - means that the chance of doing "weird stuff" that is
not directly relevant to the industry is greatly reduced.
>> I'd like to see us move towards a system
where prototype implementations of a new feature are shown to the X3D group
for possible inclusion. Right now we've been tracking a bit farther
ahead, ie needing 2 implementations by ISO approval time.
I agree we need to implement as well as specify. I
personally think the Consortium rushes too soon to ISO standardization.
I believe we should update our standards annually to enable new
capabilities - but ONLY the aspects of the standard that gain industry
acceptance get forwarded to ISO. The bar for a Consortium standard
should be two implementations - the bar for an ISO standard should be industry
acceptance.
OpenGL goes even farther than this - it insists that
radical new functionality should be battle-tested as extensions - and proven
in the industry before becoming a core part of the spec. Something
to consider.
Neil
Alan Hudson <giles@yumetech.com> Sent by: owner-www-vrml@web3d.org
Tony Parisi wrote:
> Second, there is a fallacy here about process. To think that hundreds
of
> users, authors and students you cite can *design a system* together
is quite
> naïve. It just doesn't work that way. The broader group can be deeply
> involved in defining requirements and reviewing designs all throughout
the
> process, but the actual design can only be done by a *very small*
group. In
> a perfect world, that's 1-5 people. With working groups and open process,
it
> tends to be double that size. But beyond size, that the process simply
> becomes unmanageable.
>
I've seen good results in larger groups, but I think they had a very
strong architectural foundation. The OpenGL process and OpenES process
both seem to incorporate more contributors then X3D with good results.
We are moving towards this model. At the beginnings of the X3D effort
I
felt excluded, but that was back in 98-99 when one company was basically
going to design the whole thing. Today we have about 10 active
contributors on a weekly basis. If we can move that 20-30 then we
have
the foundation to really answer some problems.
One area I'd like to move towards is a more implement before spec
procedure. In the early days of VRML it was not very possible for
a
non-browser implementor to propose a new node/feature. Without a
codebase to prototype from your really shooting blind.
The advent of several open source codes has made it possible for any
programmer to prototype a feature. Its not easy, but its doable.
I'd
like to see us move towards a system where prototype implementations of
a new feature are shown to the X3D group for possible inclusion. Right
now we've been tracking a bit farther ahead, ie needing 2
implementations by ISO approval time.
I feel we have crossed the line in terms of offering modern graphics
capabilities in an open standard. The graphics features in ammendment
1, Shaders and Advanced Textures bring X3D content authors comparable
functionality to a game engine. How quickly solid implementations
come
about will determine how far behind state of the art we are in practice.
My rambling point is I believe we can take an open standard and keep
pace with current technology. Its definately more work to document
and
debate but we get a better solution out of it.
I encourage you to read the proposals we've put together as proposed
extensions to X3D. These proposals represent various starting points
for features people have asked for in VRML/X3D. I should acknowledge
that a lot of the work on this page is not just Xj3D work, but
represents concepts and contributions from a wide range of the community
which we are trying to document and move forward:
http://www.xj3d.org/extensions/index.html
--
Alan Hudson
President: Yumetech, Inc.
www.yumetech.com/
Web3D Open Source Chair www.web3d.org/x3d/workgroups/source.html
-------------------------------------------------------------------------
for list subscription/unsubscription,
go to http://www.web3d.org/cgi-bin/public_list_signup/lwgate/listsavail.html