- From: Joris Graaumans <joris@cs.uu.nl>
- Date: Wed, 11 Jun 2003 10:44:25 +0200
- To: "Donald Spaeth" <d.spaeth@history.arts.gla.ac.uk>, <www-ql@w3.org>
I would be very happy if you could send me some samples of the historical data and some typical XSLT sheets you use for querying. If the data is suitable, I might want to use it in my experiments (with your permission of course) and check whether subjects are more productive with XSLT, XQuery or SQL/XML querying this kind of material. Actually, when I started looking for data and queries for my experiments, my advisor pointed me to a Ph.D. thesis from 1987 about the economical decline of the Dutch city Delft in the 18th century. In this thesis archived lists of housing holds are structured with a context-free grammar, entered in a computer, and queried with a special purpose query program. Unfortunately, the original data is lost, so I could only read the examples and conclude that the structure used is quite similar to SGML/XML. It would be an excellent use case for my research though. Regards, Joris Graaumans Institute of Information and Computing Sciences Utrecht University At 10:10 AM 6/9/2003 +0100, Donald Spaeth wrote: >I couldn't agree more. I've spent the last year developing stylesheets for >querying historical data using XSLT. I haven't had a chance to investigate >XQuery properly yet, so I'm not yet sure it will do what I want. Historical >source materials are good examples of semistructured data, but they have >been conventionally represented as relational tables. The challenge is to >represent the structure as it appears, but to query the data as if they >consisted of relational tables. (Representing the data as literal >relational tables with ID numbers, as in the XML Query Use Cases example, is >a less interesting way of addressing the problem.) > >Typical applications involve constructing frequencies and contingencies >tables which count particular element values, usually based upon >standardised data; and preparation of coded data in matrices for analysis >in statistical software. For example, one might count the number of rooms >per house, and see how these changed over time, or count the number of halls >with cooking equipment. (My research is based upon lists of household goods >from the seventeenth century.) All of this can be done in XSLT. > >If we set aside XSLT syntax errors (very common!), the most frequent errors >result from confusion about what the current unit of analysis is and where >one is in the hierarchy. I find myself checking and double-checking that >nodesets have the information I expect them to have. > >Best, >Donald Spaeth > >Dr Donald Spaeth >Senior Lecturer in Historical Computing >Department of History >2 University Gardens >University of Glasgow >Glasgow G12 8QQ > >tel. 0141 330 3580 >reply to: d.spaeth@history.arts.gla.ac.uk
Received on Wednesday, 11 June 2003 04:44:38 UTC