Re: [xml-dev] hypermedia affordances

[David Sheets <kosmo.zb@gmail.com>]

On Fri, Jul 6, 2012 at 12:19 PM, Len Bullard <Len.Bullard@ses-i.com> wrote:
> This is dr;tl for some of you.  Sorry... so I am trimming the reply list out of respect for your time.
>
> That is usually the problem of a one size fits all, but because I am not involved with HTML5 nor do I understand its raison d'etre, I defer to those who are and do.  Noodling is the royal road to innovation but with Google and Apple slugging it out in court, I am unsure of what will become of HTML5.
>
> XML doesn't need hypermedia affordances to succeed in the tasks for which it was specified.  Does JSON?  Now that is quite a different argument from there being a useful set of hypermedia affordances that become the norm for all hypermedia applications on the web including those using XML or MicroXML.  I would understand John's reluctance to include them because he and James have envisioned different goals.  Selah.
>
> For example, VRML97 did not use XML syntax.  It could have but XML was not ready when it was spec'd and SGML was too heavy for a real-time 3D application.  It was interoperable through the container API and that was quite limited.  It did have the concept of click and go but it worked very differently at the technical level of the control.  What it shared was http syntax which meant it could open a window with a contained resource.  On the other hand, traversing to a different part of the world was not an HTML style hyperlink.  It is a jump cut (different medium; different affordance; it inherited movie argot and style for valid reasons (objects in motion)).
>
> Then X3D was created to make it possible to use XML syntax and namespaces for real-time 3D.  It has namespaces but once again, it still interoperates with different media through the container API, eg, the browser although originally VRML worked in a VRML browser not an HTML browser and they were both web browsers.  The idea here is a web browser is not an HTML browser; it is a framework of objects that any application accessing the web can use and HTML is just one.  That is how hypermedia worked prior to the web:  it IS the operating system GUI.
>
> Now WebGL is attempting to side step that to put the 3D affordances lower in the application stack or "closer to the metal" to take advantage of late breaking display technology (hardware and firmware).  It may get an XML language on top and it may be X3D but likely not because of the personal and company politics among those writing the specs.  Life among the mammals.

Regarding WebGL, the standard reflects the procedural hardware
capabilities rather than a declarative content model. WebGL is
fundamentally about describing parallel algebraic computation rather
than 3D scenes or spaces. Any hypermedia affordances by WebGL concern
its computational domain and the declarative content of that numerical
language (WebGLSL).

To that end:

1. I have (perhaps clumsily) XML-ized the WebGL extension registry at
<http://www.khronos.org/registry/webgl/extensions/>. XML versions of
the extensions are linked from generated HTML by
link/@type='alternate' tags. HTML is generated from XML by XSLT. Atom
syndication of extension revisions are also available. This
XML-ization provides a structured format to declare WebGLSL (GPU
shading language) extension module signatures for machine consumption.

2. I have proposed the IETF registration of a WebGL shading language
media type <http://www.khronos.org/webgl/public-mailing-list/archives/1205/msg00160.html>
which was met with confusion and hostility on the public mailing list.
Khronos Group, despite claiming to mint Open Web Standards and putting
"Web" in the name of their alleged Open Web Standards, does not
actually integrate their standards with other web standards.
Additionally, some non-trivial amount of technical discussion
regarding the standard occurs on the private, members ($10k+/yr) only
list <mailto:3dweb@khronos.org>. I have still not received a response
to my media type proposal from any WebGL committee members (to my
knowledge; the committee's membership also appears to be secret).

3. I have proposed a WebGL extension that embeds the WebGL extension
namespace into the WWW's URI
<http://www.khronos.org/webgl/public-mailing-list/archives/1205/msg00135.html>
which enjoyed similar reception to the media type proposal.

4. I am presently developing a WebGLSL analysis tool, gloc
<https://github.com/ashima/gloc>, to add sorely needed modularity and
hyperlinking to the general purpose numerical WebGLSL web language
while accommodating the insane legacy lexical preprocessor. See also
<http://blog.ashimagroup.net/2012/03/15/introducing-gloc-for-open-web-shader-linking/>
and <http://www.youtube.com/watch?v=wQfvYBDJaMY#t=19m13s>. gloc is
written in OCaml and compiles into JavaScript. gloc runs on node.js as
well. gloc is BSD-licensed and a 1.1 release with a bunch of goodies
will land soon. glo 1.0 is already hyperlinked module metadata with
JSON and XML serializations.

I am concerned that Khronos and its member corporations (Google,
Mozilla ($300MM/yr from GOOG), Apple, nVidia, etc) are not good
stewards of Web standards and are, in fact, newspeaking the value of
the term "Web" to further their own short-sighted agendas. Is it evil
yet?

Your thoughts on these matters or my software or serialization formats
are warmly welcome. I have a good collection of various types of
hyper-references to declare...

Contributors are always welcome.

> A compelling reason for SGML, architectural forms and so on was/is to keep semantics from being captured **by a language or company**.  Otherwise putting it in the language
>
> A) Inhibits innovation, the treacle effect of distribution.
>
> B) Inhibits access, see the Patent Wars.   There are some long brutal threads on that topic somewhere in the W3C archives.
>
> This is why what Liam is suggesting is good advice.  Study all the examples you can find, not just HTML and find the best way to express those requirements without adopting something owned or locked up.  As Liam says you aren't likely to get more support than that from this community precisely because we do have experience.  We aren't antithetical; but that is the way to get it accepted and it may be a slow long slog.
>
> We have learned from hard experience that attempting to standardize or create outside the legal umbrella of a consortium or other organization with transparent norms guarantees capture if the affordances have value.  Again, life among the mammals.  You can't leave gold in the village square without a wall or guards.  For something as important as this if successful you absolutely must have that legal umbrella.  Otherwise it will be captured or it will die and then be resurrected under different norms, scrubbed of the names of the originators and a new army of orcs will arise.

And if the village square is a gated community already? What are
transparent norms? What if capture contradicts dogma? Can a Gorilla
feel empathy from its mountain stronghold?

> Remember what orcs are made of.  Their souls aren't happy about it.

To Mt. Doom!

David

> len
>
> -----Original Message-----
> From: Rushforth, Peter [mailto:Peter.Rushforth@NRCan-RNCan.gc.ca]
> Sent: Friday, July 06, 2012 1:42 PM
> To: Len Bullard; John Cowan; Mike Sokolov
> Cc: David Lee; Michael Kay; xml-dev@lists.xml.org
> Subject: RE: [xml-dev] hypermedia affordances
>
>> HTML5 appears ready
>> to cast off all other applications to become "the one true Ark"  On The
>> Web.  I think that makes it an easier target for sinking but....
>
> I think the problem with html is that it is trying to swallow the semantics of the world inside text/html.
>
> I think the problem with xml is that it doesn't have hypermedia affordances.  XML is complex because it needs to be in every environment where semantics are necessary.
>
> The solution is to cooperate, IMHO.  Hypermedia affordances that we can all live with would be a good start.
>
> Where is Gandalf when you need him?!
>
> Peter
>
> _______________________________________________________________________
>
> XML-DEV is a publicly archived, unmoderated list hosted by OASIS
> to support XML implementation and development. To minimize
> spam in the archives, you must subscribe before posting.
>
> [Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
> Or unsubscribe: xml-dev-unsubscribe@lists.xml.org
> subscribe: xml-dev-subscribe@lists.xml.org
> List archive: http://lists.xml.org/archives/xml-dev/
> List Guidelines: http://www.oasis-open.org/maillists/guidelines.php
>