RE: What Z39.50 clients support sorting, and how do the specify it?

Alan,

you can do sorting with our ICONE-2 client.  You can download it via page at:
http://www.crxnet.com/icone.php

It has been tested on a few servers, but as you point out, there is quite a degree of variance on how a sort can be conducted.

Rob




-----Original Message-----
From: Alan Kent [mailto:ajk@mds.rmit.edu.au]
Sent: 13 May 2003 08:54
To: ZIG
Subject: What Z39.50 clients support sorting, and how do the specify it?



Hi all,

I was wondering which clients out there support sorting via Z39.50,
and how do the clients request sorting? Do they ask for elements,
element sets, attribute lists, etc?

In particular, in the ASN.1

    SortKey ::= CHOICE {
	sortfield      [0] IMPLICIT InternationalString,
	elementSpec    [1] IMPLICIT Specification,
	sortAttributes [2] IMPLICIT SEQUENCE {
	    id     AttributeSetId,
	    list   AttributeList
	}
    }

Which do people use?

(The first choice is 'an element, element-group-tag, or alias support by
the target and denoting a set of elements associated with each record'.)

We were also considering of supporting multiple sort orders on a field.
To do this, we need to submit the ordering we want to use. We were thinking
of using an attribute for this. That is,

    Bib-1, USE(1), title(4)
    SortAttributeSet-1, SORT(2), dictionary-order(1)

versus

    Bib-1, USE(1), title(4)
    SortAttributeSet-1, SORT(2), ASCII-collating-sequence(2)

Or maybe use the new attribute architecture 'language code' attribute to
select which language collation sequence to use. (This is necessary for
sorting Unicode text for example.) Is there a better way?

Another question: SortKeySpec (which has nested in it the SortKey above)
specifies the ordering, missing value action, etc. Is there defined or
expected behaviour if an element set is specified which includes multiple
elements? (The description of 'sortfield' above implies multiplicity too.)
There are two forms of semantics I can think of:

    Identifying multiple elements to sort on means union the values then
    sort on the result as one new value.

    Identifying multiple elements means sort on the first element, then
    the second, then the third etc.

The first is useful if you have 'corporate author' and 'personal author'.
It would grab which ever author was in the record and sorted on it. So its
more useful when records would have one element or the other, but not both.

The second makes more sense when the elements are different classes
of things (such as titles and authors) rather than similar classes
(such as corporate and personal authors).

Thanks for any help.
Alan
-- 
Alan Kent (mailto:Alan.Kent@teratext.com.au, http://www.mds.rmit.edu.au/~ajk/)
Project: TeraText Technical Director (http://teratext.com.au) InQuirion Pty Ltd
Postal: Multimedia Database Systems, RMIT, GPO Box 2476V, Melbourne 3001.
Where: RMIT MDS, Bld 91, Level 3, 110 Victoria St, Carlton 3053, VIC Australia.
Phone: +61 3 9925 4114  Reception: +61 3 9925 4099  Fax: +61 3 9925 4098

Received on Tuesday, 13 May 2003 04:25:09 UTC