RE: [xml-dev] Is there a use for standardized binary XML (was RE: Microsoft FUD on binary XML...)

["Michael Rys" <mrys@microsoft.com>]

Dare is so much better with Words than I am, especially when we agree
:-).

See below for some feedback.

Best regards
Michael

> -----Original Message-----
> From: Bullard, Claude L (Len) [mailto:clbullar@ingr.com]
> Sent: Tuesday, November 25, 2003 12:19 PM
> To: Dare Obasanjo; xml-dev@lists.xml.org
> Subject: RE: [xml-dev] Is there a use for standardized binary XML (was
RE:
> Microsoft FUD on binary XML...)
> 
> I'm trying to work it out in short concise statements.  This is
> getting closer and thanks for taking the time to reply.  I agree
> with most of that, but you have to see the implications.
> 
> 1.  No binary will work well enough for all applications to
> make it worth standardizing on one.

[Michael Rys] This is our experience, yes.

> 2.  Binaries do work well for given applications and they are
> being developed.  This means that any one saying that binaries
> are bad because they bifurcate the interop problem are being
> a bit ingenuous.  They will bifurcate the interop of formats
> on the web per application type.  That will happen, and now
> it is a horse race or ye olde colonization scenario.

[Michael Rys] Not quite. Most areas that you see some form of binary XML
emerge (inside and outside of Microsoft) are normally tightly-coupled
applications. My and Dare's point is that you should not expose the
binary XML at the general interop level. If you want to expose the
binary format as an additional format, then your application and whoever
wants to talk to you will have to deal with the added complexity. But we
should not define a binary format as the W3C-blessed binary interop
format. Because this in the end will lead to some areas to claim XML
conformance while never-ever dealing with XML proper (would be kind of
like MS saying, Word's .doc format is just XML). 

The two examples that we have here are: 

1. XAML that can be compiled into something that is called BAML. This
format is like an IL for XAML's domain of defining UI and object
interaction. I am not close enough to their thinking to know how much
they want to have people program towards BAML, but I would assume that
they would like to use XAML at the interop level and be able to
fine-tune BAML over time.

2. Binary XML format for communication between SQL Server 2000 and its
OLEDB providers. This format is only sent into one direction and is not
exposed to the user (it is part of the TDS format). And interestingly
enough, the format will change in Yukon due to new data types etc. OTOH,
XML still provides you the stable interop.

> NOTE: I'm not misrepresenting the MS position here.  ALL of
> the major vendors are engaged in this.

[Michael Rys] Understood.
 
> Now it comes down to the question I asked Liam.  What will
> the outcome be of the perception that some entity, say the
> W3C, should control the development of these?
> 
> It is undeniable that development is happening.  All of the
> players are honest about that.  In the case of X3D, the
> W3DC published an RFP for one so that it would become
> the third encoding in the standard.  It doesn't get more
> straight up than that.
> 
> What will be the result of having a binary of XAML and a binary
> for SQL Server?  No big whoop because these can be said to
> be in their own application space.  But when there is a binary
> for XAML and XUL and SQL Server and Oracle, now interop is
> being compromised and the content on the web is balkanizing.

[Michael Rys] Only if people are going to start defining interop on this
level. Which will (hopefully) not happen if we have that many different
formats, since XML is the one format that is defined as the
"interoperable standard format". See it as a divide and conquer strategy
that attempts to work in favor of the XML standard.

[Michael Rys] If, OTOH, the W3C defines 5 or 6 binary "standard" format,
we certainly will get balkanization, since people will abandon XML and
tie themselves to one or two of the binary format (since it is now a
"W3C standard").

> Simply be clear about what 'local' means.  If XAML only runs
> on the MS desktop, and is not an on-the-wire format, ok. But
> if we are expected to use it on any other desktop OR must
> serve it out of an Oracle database, that may not be ok.  At
> least in the case of X3D, the consortium is developing ONE
> for all its members to use PER application type (X3D).
> 
> Is that clearer?

[Michael Rys] Is my reply clearer?

> len
> 
> 
> From: Dare Obasanjo [mailto:dareo@microsoft.com]
> 
> <len>
> So on the one hand, we have someone telling us the XML binary isn't a
> good idea for interop; on the other, we have a rich application client
> language developer telling us that is precisely what is intended.  Is
> this a fiefdom issue, or would we expect BAML to emerge as the MS
> choice for binary XML?  I'm not after MS's throat here; I simply am
> trying to show why it is so difficult to take the MS presentation
> at the binary workshop seriously.   In the rich client 3D world,
> we already take the need for the binary seriously.  It is simply
> a question of generality.  So far, all I see emerging as a consensus
> is 'the need depends on the application'.
> </len>
> 
> Is it misrepresent Microsoft's position on binary XML month? Despite
> submissions of position papers, numerous mailings to XML-DEV and
weblog
> posts it seems every other mail on XML-DEV about binary XML (or
article on
> XML.com) is about completely misrepresenting our position.
> 
> Our position has been consistent and it has been clear. Different
> applications have different optimization requirements and thus it is
> unlikely that a single binary XML standard will satisfy all scenarios
> (we're
> pretty sure it won't satisfy all the scenarios of the various
individual
> Microsoft products) given that in some cases they are conflicting.
Even it
> was the case that a single binary XML standard could somehow satisfy
all
> scenarios and not end up turning into something like W3C XML Schema
there
> is
> still the fact that this poisons the well with regards to the
> interoperability of XML on the Web. Given both these points we are
against
> standardizing on binary XML format(s).
> 
> Nothing in the above argument precludes applications from having
optimized
> representations of XML for their local needs. Does the fact that
Microsoft
> Word can accept WordprocessingML and .doc files somehow mean that .doc
> files
> should be the basis of building a binary XML standard or that you
suspect
> .doc files will emerge as Microsoft's choice for binary XML (whatever
that
> means)?
> 
> I fail to see where the inconsistency in the Microsoft position
arises.
> Len
> maybe you can explain to me why you fail to take our position
seriously?
> 
> -----------------------------------------------------------------
> The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
> initiative of OASIS <http://www.oasis-open.org>
> 
> The list archives are at http://lists.xml.org/archives/xml-dev/
> 
> To subscribe or unsubscribe from this list use the subscription
> manager: <http://lists.xml.org/ob/adm.pl>
> 
> -----------------------------------------------------------------
> The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
> initiative of OASIS <http://www.oasis-open.org>
> 
> The list archives are at http://lists.xml.org/archives/xml-dev/
> 
> To subscribe or unsubscribe from this list use the subscription
> manager: <http://lists.xml.org/ob/adm.pl>
>