- From: Sebastian Hammer <quinn@indexdata.dk>
- Date: Wed, 22 Nov 2000 22:40:49 +0100
- To: wald@library.ho.lucent.com, www-zig@w3.org
At 11:30 22-11-00 -0500, wald@library.ho.lucent.com wrote: >so why have numeric USE values? And why have attribute sets? >I have been debating this for myself - once you have EXPLAIN >and the server tells the client what attributes go with >the user selected search (e.g. "title") and these are sent back - well then >why bother? Just tell the client the allowed searchname/field names that >are searchable. And have what the user selected sent back! But I never bring >this issue up - violates all sorts of things - but since >I see the same issues appearing in explain-lite I felt it was time to point >this out. I feel that some of what confuses people when they first see Explain is that it (I believe) has the capability to both describe an attribute set, and identify the subset of a known attribute set which is supported by the server - I forget the exact category names at the moment. The latter is obviously very useful if you already know the set - you can represent the search options to the user whichever way you see fit, and use the Explain info to determine which searches (attribute combos) are actually legal. The former (decribing a set) is much more doubtful - especially if you define a set that has more than just USE values - and who's to say that USE values have attribute type 1??? Enter the new attribute set architecture to the rescue. By specifying where in the set the different things go, it seems to me we have a much better chance of making things work... but enter semantic qualifiers and how to describe them... sigh. Someday we'll get there. Maybe part of the reasy Explain uptake is slow is that for the vast majority of cases, we can do business without it? After all, most deployed Z39.50 clients are probably WWW gateways operated by various types of hosts and for various purposes. They tend to provide you with a set list of resources which have been deemed relevant by the host (which is generally a Good Thing). You don't generally "surf" amongst Z39.50 servers the way you do web-servers. One could postulate (thinking out loud here) that this may be the natural order of things, since the purpose Z39.50 serves is normally pretty specialised. In a closely profiled, closed community - which is something we see more and more in the library world with the uprising of detailed national profile - you don't really NEED Explain because you already know the important things about the server - what attribute combos and record syntaxes, etc. are supported. You don't know what databases the server offers, but for the most part you don't want to - you mostly want to hit the database that corresponds to the OPAC function - not all the junk databases around it, and this you pick out of the Z39.50 URL. Again - just thinking out loud. Certainly, in Danish library circles, we seem to attach a higher priority to things like holdings info, item ordering, possibly even scan and sort, than we do to Explain. --Sebastian -- Sebastian Hammer <quinn@indexdata.dk> Index Data ApS Ph.: +45 3341 0100 <http://www.indexdata.dk> Fax: +45 3341 0101
Received on Wednesday, 22 November 2000 16:41:13 UTC