Re: [Fwd: Implementation of XML QUERY]

[Scott Hinkelman <srh@us.ibm.com>]

David W. has good points here about separation of actors concerning query;
the general public exposure
for registry contact is likely most appropriate for business-level
interfacing, while power-user flexibility, as
in adhoc, is in a separate space of use. This somewhat gets back to the 'a
DBA perspective' thread. It is difficult to cast
an interface that explicitly outlines adhoc query specification that does
not suffer from implementation-seep.
I believe it is likely a good move to think toward a business-level
interface [list_CoreComponents(x,y,z)] for
these reasons and for reasons of ebXML business-level buy-in.

Scott Hinkelman, Senior Software Engineer
XML Industry Enablement
IBM e-business Standards Strategy
512-823-8097 (TL 793-8097) (Cell: 512-940-0519)
srh@us.ibm.com, Fax: 512-838-1074



David RR Webber <Gnosis_@compuserve.com>@compuserve.com> on 01/23/2001
09:45:13 PM

To:   Farrukh Najmi <Farrukh.Najmi@east.sun.com>
cc:   "ebxml-regrep@lists.ebxml.org" <ebxml-regrep@lists.ebxml.org>,
      chetan@pybiz.com
Subject:  [Fwd: Implementation of XML QUERY]



Message text written by Farrukh Najmi
>
FYI,

In my quest for query experts and their opinions I came across Chetan on
the xml-dist-app list. Coincidently, it seems that they are discussing
this issue of late.
<<<<<<<<<<<<<<<

Farrukh,

It has come to my attention that you are flogging a dead horse here.

Why are we scouring the globe for 'query experts' - whatever they
may be?  (Will the REAL query experts please stand up - spotlights,
drum roll, applause).

Frankly, brutally, obviously - this whole debate of foobar query X vis
barfoo query Y is totally and completely IRRELEVANT.

Can we please stop this now?

The concensus I saw was that we should be describing our
interfacing in neutral terms, that match the process of discovering
information via the RIM in simple english language documentation,
then map those to a set of descrete methods that we publish and
following that implementors can pick whatever backend
datastorage system and querying to make that happen.

We have one consistent set of methods for the service API
that a registry exposes.

Now - Mark Hale pointed out on the TC the other week that
his experimenting with a UDDI registry that follows this
approach - that this was not enough for him - he needs true
ad hoc.  Notice Mark is a super-user, notice Mark is using
ad hoc to get at all sorts of IMPLEMENTATION specific
backend details - cos its his job to be a programmer.

That is the whole point.  Programmer = backdoor access to
internals of registry - probably on a dedicated port via a
firewall - internal access.

Normal business user - descrete access via industry functional
suite of services and API - via public URL and open port.

Can we please agree these are two totally different uses.
That programmers will use ad hoc query in whatever internal
query syntax the particular backend implementation
provides.

Public users will have the ability to use a set of RIM service
API methods that provide "lego bricks" to assemble
constrainted querying - and that these will NOT use a
query syntax but instead query syntax agnostic methods
that provide function + value parameters.

These will also allow static business querying of content
via business parameter values.

This then is not rocket science - but business basic
analysis work.

BTW - since I've taught University level SQL and have been
writing database backend code for twenty years
and worked on Oracle pre-processor for C and COBOL,
and worked on Assembler level interface to ADABAS,
and currently work for a leading vendor of XML searching
and querying tools, and have a patent for information
manipulation techniques, and currently on W3C working
groups.  How come I'm not one of your 'query experts'?

Can you publish your criteria for what is a 'query expert'?

Oh - yes - and can we please ditch trying to use SQL-78
for implementing XML based registry - well you can if
you like - but the rest of the world would rather not be
burdened with this - its rather like telling Frank
Whittle to design a jet aircraft, and then telling him -
oh and BTW - the production system will actually
use a piston engine - because that's what our
mechanics know how to work on right now.

Respectfully pointing out the obvious in the hope of
real progress here.

Thanks, DW.