Re: [xml-dev] A question about the future of efficient XML

["Stephen D. Williams" <sdw@lig.net>]

Richard Salz wrote:
>> I think the point you are missing here is that an efficient XML 
>>     
> interchange 
>   
>> technology is not intended to "replace" XML, it is intended to 
>>     
I would say it is not _primarily intended_ to replace XML.  Primarily, 
it allows XML technology to be used in applications where XML is 
currently too inefficient, in size or processing speed, _and_ to improve 
those cases where it is a significant overhead cost.  More and more 
cases fall into that last category, meaning that efficient XML will, 
IMHO, likely have the effect of augmenting XML gradually even for many 
current applications.
> "complement" 
>
> The difference between "an efficient XML interchange" and "the efficient 
> XML interchange" is lost on many people.
>   
There are good, general, techniques for much of what needs to be done to 
create "the" efficient XML interchange.  The XBC and EXI groups examined 
many different ways to meet the requirements, after hammering those 
requirements out and carefully ranking them.  The initial spec is, I 
think, a baseline that forms a good foundation.  It opts for minimal 
features and relative simplicity while greatly improving nearly all use 
cases.  Techniques are selected that significantly outperform other 
methods overall.  Some want to stop at the baseline, others of us see 
some significant improvements through certain changes or additional 
features which we will shortly be in a position to test and prove.

One of the great things about this work is the clarity of model for data 
exchange and representation in terms of requirements, measurement, and 
the concept of a spectrum of data and metadata that ranges from 
self-contained instances to "fully externalized" instances.  This model, 
in it's fully realized form, unifies the spectrum from a very XML-like 
fully self-contained instance with no schema to a PER-like set of a few 
bits, all represented and interpreted in the same way.

> *MY PERSONAL VIEW* is that it's a good Occam's Razor for analyzing the 
> membership of the W3C binary XML WG.
>   
And analyzing the non-membership of the group is a good Occam's Razor 
for XML vs. efficient XML.

sdw
>         /r$
>
> --
> STSM, Senior Security Architect
> DataPower SOA Appliances
> http://www.ibm.com/software/integration/datapower/
>
>