RE: Andrew Layman and Don Box Analysis of XML Optimization Techni ques

Len Bullard wrote:

> Is this a disruptive technology?  Likely yes.
> Incumbents never like those and do whatever they
> can from proposing overloaded requirements, humor
> pieces designed by showmen to muddy discussions,
> FUD, whatever to slow those down until they have
> their own products ready.

Well, I think we should tone down this whole discussion.  As I am partly 
responsible for heating it up a bit, I apologize for that.  I do think the 
above somewhat oversimplifies the reasons for my concern.

I suspect you meant "disruptive" in the Clayton Christensen sense [1] of a 
technology that percolates up from the bottom, is initially not fully 
robust or general, but is compelling for certain applications.   Such 
innovations often mature to overwhelm established technologies and along 
with them, established business models of "incumbents".  Fair enough.

My concern is that Binary XML is disruptive in another less positive 
sense.  Part of the value of XML is its nearly universal interoperability. 
 XML data can be repurposed over and over again, sometimes for uses not 
originally anticipated.  You can take most any XML and read it into Excel, 
import it into a variety of databases, transform it with widely available 
XSL tools, etc.  While in principle one could re-release all the software 
that's already out there to include new drivers for binary XML, in 
practice there will for years be software that only understands the text 
form.  Even if binary is successful, we will bear for the indefinite 
future the cost of conversion between the two, e.g. when editing in Emacs 
is desired.  So, there is a downside.

I'm not personally against Binary XML.  I do think the disruptions in the 
2nd sense are sufficiently troublesome to those who benefit from XML today 
that we should set the bar fairly high in justifying Binary XML. 
Reasonable people can disagree over whether the analysis done by the 
Characterization WG gets us over that bar.  I don't think my concern is 
driven primarily by my employer's role as an "incumbent".  I won't be 
surprised if a careful analysis shows that there is indeed compelling 
value in Binary XML for some of the use cases that are on the table, and 
that we should indeed do a Recommendation.  I'm just not yet convinced 
that we know which use cases we can really address, and what factors in 
space or time to expect in return for the investment.  I'd like to know 
that before signing on.

And, FWIW, I think that any time we have an opportunity to introduce the 
Navier-Stokes equations into the design of XML, we should leap at the 
chance. 

Noah


[1] 
http://www.amazon.com/exec/obidos/ASIN/0060521996/qid=1112892395/sr=2-1/ref=pd_bbs_b_2_1/104-9643018-5330327

--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------

Received on Thursday, 7 April 2005 17:00:22 UTC