- From: Matthew Dovey <matthew.dovey@las.ox.ac.uk>
- Date: Mon, 22 Apr 2002 14:13:12 +0100
- To: "Sebastian Hammer" <quinn@indexdata.dk>, "Robert Sanderson" <azaroth@liverpool.ac.uk>, "Theo van Veen" <Theo.vanVeen@kb.nl>
- Cc: <www-zig@w3.org>
> Another rationale is that the retrieval records may be very large. If > you've found ten books in a full-text base and you want to show the titles > to the user, it would be a shame to have to download ten complete books > from the server. There is a another real-life example which comes out of the holdings schema for union catalogues. This lead to a eSpec-Q. The problem here is that you are searching a union catalogue of say 120 different libraries (as in the situation in Oxford, but is equally applicable to national union catalogues etc., I think eSpec-Q came out of the French union catalogue). Now in this situation, for a fairly common book you may get back 120 different locations. For a typical user, only 1 or 2 specified libraries are going to be relevant (based on which libraries you can borrow from or which are physically close). So you really want to say to the server - give me the bibliographic record and the holdings statements for the following 2 or 3 libraries. Pulling back all 120 holdings statements and then filtering client side seems inefficient of bandwidth. This gets worse if say you are using a European Union Catalogue (1000s of libraries) over a modem connection... Matthew
Received on Monday, 22 April 2002 09:13:13 UTC