W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > July to September 2004

Some thoughts / thanks for Carlsbad attendance

From: John Bresnik <brez@brezbrez.com>
Date: Wed, 21 Jul 2004 11:59:11 -0400 (EDT)
Message-ID: <20040721054217.50569.qmail@web14801.mail.yahoo.com>
To: public-rdf-dawg@w3.org

This is my second attempt at posting, so disregard the
redundancy if applicable. Anyhow, thanks for letting
me attend.  It was my first inside look at a W3C
process and a very interesting experience.

Some thoughts: 

thought 1. 

As far as the many flavors, everything seemed to come
back to a question of *how a graph is being stored [is
it in RDF, N3, RDBMS, etc] - Since there are
potentially infinite ways to abstract a graph then, it
seems, we need a language that deals with them in the
abstraction of a graph, i.e. letting *how the graph is
stored drive how it is queried is fraught with
problems of scalability, adaptation to new ways to
store graphs, etc. 

thought 2. 

>From an N-tier architecture perspective, the fact that
a query language has conditional logic makes me cringe
- mainly because the temptation to violate N-tier
precepts [embedding business logic in your data access
layer] is too great. [cf. mis-use of triggers / stored
procs in enterprise RDBMS]  

thought 3. 

Re. the cursors, etc. discussion [requirements] – I
wouldn’t underestimate the reality that a caching
mechanism [like those found in the object-relational
mapping tools like Toplink / ext to Hibernate] is on
the horizon, i.e. perhaps err on the side of *less
features for now – particularly those that are prime
candidates for a caching mechanism [like scrolling a
result set/book-marking, etc.].  

thought 4. 

I think trying to compromise the divide by simply
developing a SQL-like/X-Query composite language
doesn’t take into account a new, as yet undeveloped,
way to store graphs.  ? or all the existing ones for
that matter. [see note about storage-driven decision
making above]


Andy spoke of a second phase of the process where a
number of implementations are done [ideally by
outsiders] - ? I'd like more information on that if
someone gets a chance - i.e. it's an area that we
could be of some use since we'll soon be in that
space. Also I can sort out membership info via the
NAVY if thats a req. 

It was a pleasure meeting all of you. 

Thanks again,
John Bresnik 
SPAWAR Systems Center, San Diego. 
Received on Wednesday, 21 July 2004 14:37:33 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:00:28 UTC