Re: [xml-dev] XML aggregation question?

[ht@xxxxxxxxxxxx (Henry S. Thompson)]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Andrew S. Townley writes:

> This is a valid point.  I know it will make a difference, but I'm trying
> to evaluate it on the smallish scale:  lots (1000s) of small docs in the
> < 1K range which might have a lifespan of a week to years.  The
> reliability of the data is important but the availability isn't quite as
> crucial so long as it is reasonably available for some vague definition
> of "reasonably".  :)
>
> At the moment, I'm still in the thinking/prototyping phase to see if my
> idea is even viable or not.  If it is, it could evolve into a couple
> orders of magnitude bigger and with appropriately more demanding
> availability requirements, but that's waaay down the road, but it is a
> consideration.

Well, for the phase you're in, and even the next phase, I'd encourage
you to look at an all-XML no-DB solution.  The kind of architecture
I've used for e.g. an admissions management application (1000
applications, 1.5K doc per applicant) works with a master document
like this:

<decisions xmlns:xi="http://www.w3.org/2001/XInclude";>
 <decision>
  <xi:include href="a34034.xml"/> <!-- an <applicant> element -->
  <rank>32</rank>
  <outcome>Accept</outcome>
 </decision>
 <decision>
  <xi:include href="a34182.xml"/> <!-- an <applicant> element -->
  <rank>50</rank>
  <outcome>Pend</outcome>
 </decision>
 . . .
</decisions>

Convert this for your purposes into
 <statuses>
  <status value="xxx">
  <xi:include href="w34034.xml"/> <!-- a <widget> element -->
  </status>
  . . .

Plus pipelines [1] which e.g. run XInclude followed by XSLT to extract
appropriate views.

Obviously way too schematic, but gives you the idea, I hope.

Or you may want to have two layers of virtual documents, with one file
per status, and 

 ( echo "<statuses>" ; cat currentStatus/*.xml ; echo "</statuses>" ) | \
  [XIncl and XSLT pipeline]

ht

[1] http://www.markup.co.uk/
[available for design consultancy contract :-]
- -- 
 Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
                     Half-time member of W3C Team
    2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
            Fax: (44) 131 650-4587, e-mail: ht@inf.ed.ac.uk
                   URL: http://www.ltg.ed.ac.uk/~ht/
[mail really from me _always_ has this .sig -- mail without it is forged spam]
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.1 (GNU/Linux)

iD8DBQFE8tDwkjnJixAXWBoRAn7zAJ9sTQUVBhUY3JNj6eQE5KdsNbqdrQCffBwT
W/xKKBjIts2+6vvleyDVgL4=
=NlOJ
-----END PGP SIGNATURE-----