XQuery and XPath Full-Text Use Cases

Hi Vagelis,

Thanks for your comments on the XQuery and XPath Full-Text Use Cases.

My answers are inlined.

Also, for further comments, could you please use this email address:
public-qt-comments@w3.org

It will read the whole FTTF (full-text task force) group and you might
get answers from people with different perspectives (industry,
information retrieval, librarians, ...).

I am creating a thread in the above mailing list in case people have
more comments.

Thanks

Sihem


>From: "Vagelis Hristidis" <vagelis@cs.ucsd.edu>
>Cc: "Yannis Papakonstantinou" <yannis@cs.ucsd.edu>
>Date: Mon, 17 Mar 2003 14:02:44 -0800
>X-Priority: 3
>X-MSMail-Priority: Normal
>X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106
>X-MailScanner: PASSED (v1.2.7 40182 h2HM15KY071993 mailbox6.ucsd.edu)
>X-Spam-Status: No, hits=-6.6 required=5.0
>	tests=BAYES_01
>	version=2.50
>X-Spam-Level: 
>X-Spam-Checker-Version: SpamAssassin 2.50 (1.173-2003-02-20-exp)
>Content-Type: text/plain;
>	charset="iso-8859-1"
>Content-Length: 965
>
>Hello Sihem,
>we are looking into the problem of keyword search in XML databases (our most
>recent work is Vagelis Hristidis, Yannis Papakonstantinou, Andrey Balmin:
>Keyword Proximity Search on XML Graphs. ICDE, 2003), and while reading the
>W3C draft I got a few questions:
>

Yes, indeed. I wanted to come listen to you at ICDE but I had a
conflict at that time. I will definitely read your paper.

>1. Which is the difference between cases 13.2.3 and 13.2.5? They both seem
>to look for two phrases where the first one is in an ancestor and the second
>in a descendant node
>

There is not always a major difference between queries. In particular,
the two queries you refer to are in Section 13: Use Case "STRUCTURE":
Queries using XPath Axes whose aim is to combine XPath navigation with
full-text querying and the only difference between 13.2.3 and 13.2.5 s
that one uses the Descendant axis while the other use the Ancestor
one.

>2. Do you support plain keyword queries of the form: give me an element with
>its subelements, such that they contain all the keywords of the query? That
>is, the user does not specify any element types, but only a set of keywords.
>

All queries seem to specify exactly where to search for keywords. We
do not have a use case that explicitly deals with your concern. It is
definitely a use case we will have to consider. To make sure I
understand your request, an example of the kind of queries you would
like to see is:

Return book titles and authors for books that contain the words
"usability" and the word "performance" (meaning anywhere in their
descendant elements without specifying which elements they are).

>3. How about IDREFs? Do you plan to support phrases spanning across IDREFs?
>

We do not have any plans to support searching for phrases spanning
IDREFs. Do you have any example we could discuss?

>Thank you for your time.
>
>Regards,
>Vagelis
>
>
>--
>Vagelis Hristidis
>Computer Science & Engineering Dept
>University of California, San Diego
>email: vagelis@cs.ucsd.edu
>URL: http://www.db.ucsd.edu/people/vagelis/
>
>

Received on Thursday, 20 March 2003 14:37:19 UTC