RE: Proposed XQuery requirement and/or objective

Jim-

 

Points taken, and no hostility inferred.

 

Your counterpoints regarding the adoption of SQL are a great debate to
have.

 

In broad brush-strokes, we are committed to a query concrete syntax
which is grounded in a widely-adopted (and preferably W3C recommended)
representation.

 

Further, in no means do I intend to imply that XQuery would make things
easier on the vendor implementations for RDFS/OWL/Rule components of the
SemWeb - quite the opposite, the implementations may even be more
difficult.  Our point is intended to speak towards our opinion that a
known query representation would speed user adoption rates for semantic
web languages.

 

If early adopters of large commercial organizations were faced with
learning and implementing a wholly new syntax for queries - on top of
what they already have to pay for in human resource expertise - we
suspect, and have encountered, resistance.

 

Anecdotally, we would likely be supportive of the OWL "two surface
realizations" model, as long as one of them was a widely-adopted
standard format.

 

-Jeff-

 

  _____  

From: Jim Hendler [mailto:hendler@cs.umd.edu] 
Sent: Thursday, July 15, 2004 1:29 PM
To: Jeff Pollock; public-rdf-dawg@w3.org
Subject: Re: Proposed XQuery requirement and/or objective

 

At 11:24 -0700 7/15/04, Jeff Pollock wrote:

All-

 

As everyone in attendance at the DAWG FTF is aware, the Network
Inference position is that XQuery represents an ideal concrete syntax to
facilitate more widespread and rapid adoption of the semantic web stack,
beyond just RDF.

 

In support of the charter, NI proposes that the discussion of a strawman
query language is not inherently at odds with an eventual XQuery based
concrete syntax. The DAWG charter contemplates three stages of
specification evolution; (1) strawman language, (2) abstract syntax and
(3) concrete syntax. Network Inference does not perceive BRQL, RDQL, or
N3QL proposals as necessarily excluding an eventual outcome that is
consistent with an XQuery concrete syntax. The DAWG charter also clearly
states that: "The RDF DAWG should aim to maximize W3C technology re-use,
while also taking account of differences between the RDF graph data
model and the XQuery data model" - thereby supporting our objective that
XQuery be an integral part of an end-user's interface with the eventual
DAWG language recommendation.

 

Therefore, NI proposes that a new requirement be considered by this
group:

 

"The query language shall have an XQuery compatible concrete language
syntax."

 

Network Inference has formalized the rationale for this requirement in
the attached document that explores the facets of why this is important
to us, and ultimately as to why it should be important to the W3C
community at large.

 

Comments, stones, or support are welcome.

 

Kind Regards,

 

-Jeff-

 

 

 

Jeff - thanks for this document - it seems helpful - since I'm not there
I cannot really follow the arguments, but I must admit I find this
document confuses me more, not less.

 

some questions;

 you say 20+ vendors support some sort of Xquery.  best I can tell,
however, most of those assume you are using a XML DB.   Some 200+
vendors seems to support SQL, including MS access and Oracle and lots of
the other "big kids" -- and they don't assume an XML DB.  

 I'm assuming that in most of the applications people will build, the
RDF will live in a triple store (like Tucana builds, for example) and
thus they won't be in an XML DB per se -- so would Xquery really speed
up adoption or slow it down?  SQL is also taught in every CS dept in
America (I can't speak for outside the US, although I did sit in on some
DB lectures using SQL in the UK, so it's at least taught there as well).
There are also many packages (like JDBC) that provide APIs for working
with SQL-like langauges, wouldn't that be an argument for that sort of
syntax?

 

You show RDFS/OWL/Rule query langauges as somehow being more easy in
Xquery, but again I think that is because you are assuming these things
will be kept in their RDF/XML documents, or in APIs that respect the
"boundaries" of those.  I already see many applications moving towards
multiontologies w/linking, and that seems to me to argue that we simply
don't know yet which of these models are better.

 

Let me be clear, I'm not opposed to an XQuery-based model, and I am not
really a big fan of SQL, but I see a lot of the DB community folks
(Several papers at TODS and VLDB workshops, in fact) that are arguing in
favor of RDF DBs precisely because they can handle algebras similar to
relational, and because they can scale well in the way RDBMS do (and
they're worried about the scaling of XML DBs, esp in the VLDB world)

 

So I would be okay with the requirement you propose assuming

 one - this can't be the ONLY concrete language syntax - I would expect
that we could do something like we did in the OWL case, where we had an
abstract syntax with two surface realizations (XML and RDF/XML) - this
guaranteed that interoperability between the XML and RDF version was
possible, as they were provably mappable one to the other

 two - I would like to see some evidence of the scalability of these
Xquery things -- this is not meant to be hostile, I'm just a lot more
comfortable with the SQL world (where I have over 25 years of
experience) than with the newer Xquery world, so I simply don't know
what Xquery Algebras and calculi look like, and thus I don't have an
inuitive feel for their scalability, etc.    I suspect Tucana and others
would be willing to make their DB query times public for some set of
queries - is there any way we could run a comparison on some large scale
test cases or someting like that (I looked on the NI site for some
examples, but there is nothing publicly available)

 

Please, let me be clear -- rereading this it looks like I'm being
negative, but I don't mean to be -- I'm actually trying to understand
the situation and see whether the positions we seem to hold are
compatible or not -- I certainly have not predecided this issue

 -JH

p.s. And if you think I may be somehow hostile to NI, please talk to
Jack Berkowitz and get him to tell you how supportive I was during the
Web Ont work.

-- 

Professor James Hendler
http://www.cs.umd.edu/users/hendler 
Director, Semantic Web and Agent Technologies       301-405-2696
Maryland Information and Network Dynamics Lab.      301-405-6707 (Fax)
Univ of Maryland, College Park, MD 20742      240-277-3388 (Cell)