Re: binary XML

[redirecting: www-tag@w3.org, "w3c-tag" doesn't exist.]

[cc: CAD3D, X3D-Spec teams, Rob Glidden and Joseph Reagle]

Paul, thanks for the detailed message.  Here are details in reply.

Paul Cotton writes:
> Near the end of the AC meeting this week you mentioned that your users
> in the Web 3D Consortium need a "compressed XML" format and that you
> would like the W3C to work on this (from IRC log at
> http://www.w3.org/2002/11/20-ac-irc ):
>
> 22:24:21 [RalphSw] [scribe throughout]
> Don Brutzman:
> ... Web3D brought in 30 members on CAD mapping for the Web
> ... mapping has not yet gotten to XML, but that seems an obvious thing
> to do
> ... SVG or other ways to map CAD would be good to do
> ... 30 members joining overnight is an unusual event; a couple are
> already W3C Members
> ... technical point: we will need a compressed XML.
> ... we're not keen to invent it ourselves
> ... we ask that W3C take a look at this
> Joseph Reagle:
> ... Don and I talked earlier and discussed encryption of this data
> ... I think we could accomodate this, but it is not something we took on

Background:

- new CAD3D Group in Web3D Consortium is currently examining
  requirements for mapping CAD models to the Web.  Group
  motivation is to enable technical cooperation between vendor
  companies to help resolve the issues of sharing and viewing
  of CAD/3D data via a common file format & viewing engine.
	http://www.web3d.org/fs_cad.htm

- Extensible 3D (X3D) Graphics expects to add a binary format
  as an alternative encoding, with requirements excerpted here:

=====================
The X3D group recommends consideration of technical requirements in the
original X3D Binary Compression Requirements and Request for Proposals,
online at http://www.web3d.org/TaskGroups/x3d/binary/X3dBinaryRFP.html
Editorial credit:  Rob Glidden authored the original binary RFP.

Now that we are completing the X3D Specification, composition of
geometric compression and binary encodings (along with encryption
& authentication) stands as the next major area of new work for X3D.
Interestingly, XML encryption and authentication (with pluggably
replacable algorithms) is already available via W3C Recommendations,
though binary and geometric compression are not.  Once an aggregated
extensible solution is available, we expect that this work will
provide a major new section in our ISO X3D Specification.

We believe that the nine separate RFP areas remain closely interrelated
and mutually achievable.  X3D binary compression requirements include:
1) Multiple data types (Geometry, Interpolation data, Textures,
   Media, Tokens)
2) Processing Performance
3) X3D Compatibility
4) Interoperability
5) Ease of Implementation
6) Streaming
7) Authorability
8) Compression
9) Security

Metadata issues likely need to be called out as an additional area,
or possibly treated as a special data type.
[...]
Of note, X3D extensibility offers particular opportunities for integrated
encoding of specialty CAD information that is selected for various Web
applications.  Prototype declaration and instancing allows creation
of customized, repeatable scene-graph node templates.  Scripting via
EcmaScript or Java allows programmatic access and creation of data,
as well as Web updates of 3D scene information.  Exchange of event
messages via import/export connections across different scene files
maintains animation and interactivity even when building shared scenes
from archived model libraries.  Finally the use of multiple XML
namespaces in a single X3D document can allow close and validatable
integration with server-side databases or CAD-centric custom metadata.
=====================

One possible shared path for these related efforts to take might be
- representation of CAD files in extended X3D XML-encoded documents
- various geometric-compression algorithms on geometric payload data
- XML encryption, authentication via W3C recommendations
- XML compression/decompression of result

Finally, if we start with all the goodness of XML for 3D scene data,
there are two compelling reasons to have XML compression for 3D scenes:
- increased efficiency, size reduction due to highly structured documents
- increased parse speed, accelerated by a priori tokenization

So I think it likely that one or both Web3D endeavors will need
XML compression.

> 22:27:09 [DanC]
> hmm... isn't binary infoset a TAG issue? I don't see it in
> http://www.w3.org/2001/tag/ilist ...
> 22:27:51 [DanC]
> not accepted as an issue yet, but: # Binary Infosets Robin Berjon (Wed,
> Oct 09 2002)
> http://lists.w3.org/Archives/Public/www-tag/2002Oct/0187.html
>
> You should note that Dan Connolly (Team member and W3C TAG member)
> supplied a pointer in the IRC log to a message providing input on the
> text in the W3C  Architecture document on "binary infosets".  This email
> points to an attachment on the author's work on "binary XML" which you
> might be interested in reviewing.

yes, very interesting

	Bin-XML™ for encoding XML documents
	http://expway.tv/graph/Bin-XMLTechnical%20White%20Paper.pdf

another common reference on this subject is
	XMILL:  An Efficient Compressor for XML
	http://www.research.att.com/sw/tools/xmill

> But in addition I should point out Tim Bray's response to the above
> email
> http://lists.w3.org/Archives/Public/www-tag/2002Oct/0188.html
> in which he stated:
>
> <personal-opinion>Binary infoset?  Blecch.  This is what XML exists to
> prevent.</personal-opinion>

having just read that thread, I don't see much value in a separate
binary infoset, but can see need for a binary encoding that supports
XML Infoset.  Open issues include whether an encoding for a given
schema might be fixed, reducing the need to transmit encoding info
and enhancing optimizability of schema-specific binary parsers.

> Given that you think "compressed XML" is a W3D requirement I invite you
> to join the www-tag@w3.org debate on this matter by possibly replying to
> one of the above messages.

thanks, can do.  will extract/post CAD3D-X3D excerpts from this message
following any comment.  our next CAD3D meeting is late January.  I'll
request a discussion period then, and will find out which participants
are also W3C members who might want to lend their support for such an
effort.

given possible interactions with encryption/authentication, the likely
need for compression by a variety of XML languages/schemas, and
long-term brittleness of ad hoc solutions, XML binary compression will
benefit from consideration by W3C.   am glad you flagged it as an
architectural issue that hopefully will deserve some TAG attention.

> /paulc
> W3C TAG member
>
> Paul Cotton, Microsoft Canada
> 17 Eleanor Drive, Nepean, Ontario K2E 6A3
> Tel: (613) 225-5445 Fax: (425) 936-7329
> <mailto:pcotton@microsoft.com>

again thanks Paul.

all the best, Don
--
Don Brutzman  Naval Postgraduate School, Code UW/Br Root 200  work
 831.656.2149 Monterey California 93943-5000 USA              fax 
 831.656.3679 Virtual worlds/underwater robots/Internet    
 http://web.nps.navy.mil/~brutzman

Received on Tuesday, 26 November 2002 10:25:04 UTC