A 9:06 +0000 2/12/97, Dan Brickley a écrit:
>On Tue, 2 Dec 1997, Andrew Waugh wrote:
>
>> Stu wrote...
>> > I would recommend that you follow standard proceedures that you would
>> > employ for other cataloging activities ... so, for names, I should
>> > think Last name, first name, middle initial would be most desirable.
>[...]
>> My bad old days in electronic directories taught me that the best
>> stored name is the name preferred by the person. This causes the least
>> amount of havoc when you are suddenly presented with an unusual name
>> (especially if the person concerned is sensitive about their name).
>> This is particularly true when you consider that the order of the name
>> components need make no difference to searching in DC.
> ^^^^^^^^^^^^^^^^^^^
>
>It can make a vast difference to a browsable display of DC records though.
>DC is oriented towards resource 'discovery'; this isn't always the same
>thing as searching. If we treat searching as the only application of DC,
>we risk creating records that are hard to display as, for eg,
>alphabetically ordered lists. I've no argument with your main point
>though - names are a mess. Just that we shouldn't forget the implications
>for browsing when we decide how to work around this mess.
Exactly.
Right now, an indexer will usually index words, and, given you know how to
cheat it into giving you what you want, this is ok. But I do display content
of dc records, and so I would like them to look like one another if possible.
Also it would be nice to be able to click on what is displayed (ie the name)
(for instance to get a little less noise), therefore it would be easier for
the re-searching process to be able to re-generate a pertinent question, etc...
Maybe one day we write (or buy, or get or whatever) indexers which know
about more than words, and then the question of separators gets to be important
too.
However, one thing is to change the format of your metadata (this is easily
done by global changing) another is to change the data people have actually
typed in (personal names for instance).
Cheers,
[log in to unmask] Tél: 04 76 51 48 72
Fax: 04 76 63 56 11
_______________________________________________________________________
Cellule de Coordination Documentaire Nationale pour les Mathématiques
UJF - MENRT - CNRS * UMS 5638
Adresse postale : Cellule MathDoc - UMS 5638, Université Joseph Fourier
B.P. 53, 38051 Grenoble Cedex 9, France.
Serveur: http://www-mathdoc.ujf-grenoble.fr/
_______________________________________________________________________
|