Hugh's analysis of URLs in 670 reminds me a bit of the statement of 
responsibility discussion on the MARC list. Yes, statements of 
responsibility are indispensible when establishing a name or uniform 
title (including series) or when doing database cleanup (especially 
trying to differentiate shared real/pseudonym persons and earlier/later 
corporate names). To the user, statements of responsibility are not 
very helpful, particularly if they're not indexed.

On to URLs. When you're establishing a heading, a website can be real 
handy in identifying the nature of a body or how it names (brands) 
itself, at least at the moment.

In both cases though, the bibliographic or authority records cannot 
entirely substitute for a new analysis of the situation. With our 
friends Google and other browsers, that new analysis can be done when 
we need to do it. A 670 with a solid explanation of the website where 
you found the information may be more helpful as a guide to a browser 
search than a URL gone bad. That is, it is better in the long term to 
say "Comune di Vicenza WWW home page" 
than "http://www.comune.vicenza.it/"

Yes, it is Friday. And, Hugh, thanks for the URL analysis.

Sherman Clarke
NYU Libraries
[log in to unmask]

----- Original Message -----
From: Hugh Taylor <[log in to unmask]>
Date: Friday, March 9, 2007 7:28 am
Subject: [PCCLIST] More on subfield u in authority record 670 fields

> Back in September, I summarised the results of a fairly basic 
> analysis 
> I'd undertaken on the use of 670/u in authority records. I've just 
> repeated that analysis, and I'm no more encouraged by what I've 
> discovered now than I was back in September.
> 
> Our Voyager system found 1452 instances of URL-type data in 670/a.
> 
> Of these, a manual analysis suggests:
>   91 have parentheses wrapped around the URL
>   94 have the URL followed by date information (almost certainly 
> intended to be the "date viewed")
>   17 have the URL followed by other sorts of extraneous non-URL data
> 
> Some browsers will be able to make some sense of all three types, 
> but I 
> don't think there's any doubt that all are incorrect expressions 
> of the 
> "URL". So that's 202 out of 1452 occurrences of the subfield that 
> are 
> deficient in some way. Just under 14% - or one record in seven.
> 
> Actually, the figure is higher than that because I didn't include 
> in the 
> above count the typos, invalid or impossible characters and other 
> odd 
> bits of detritus that I've found in the remaining 1250. (I 
> particularly 
> liked the record in which the URL included two blanks - itself not 
> permitted according to my understanding of things - followed by a 
> non-spacing underscore... Interesting!)
> 
> Last time I wrote, I asked if anyone wished to defend their right 
> to 
> include parentheses as "wrappers" in the 670/u, or to justify the 
> inclusion of additional text? Nobody did - the messages I received 
> supported my understanding of what was "correct". And yet the 
> numbers 
> continue to increase.
> 
> At the same time, it might also be worth considering just what's 
> going 
> into other subfields in 670s where there's a subfield u present. 
> Often 
> there's no subfield b (bearing in mind these are quite recent 
> records, 
> that seems contrary to what I understood we were supposed to be 
> doing). 
> And some of the information that's going into the subfield a looks 
> a 
> little odd, or is expressed strangely, e.g.
>   main menu (Rubens Maciel; d. Aug. 24, 2004)
> 
> Finally, I threw all of the URLs at a link checker (Zenu). Here's 
> a 
> summary of the results:
>   ok                                           1040
>   forbidden request (but URL "looks" ok)         24
>   moved (either permanently or temporarily)      60
> 
>   no such host, or page not found               288
>    (NB: this includes the 91 in parentheses)
> 
> The remainder comprise a selection of technical messages (server 
> error 
> and the like).
> 
> Perhaps I just don't have enough things in my life to worry about, 
> but I 
> find quite a bit in what's above that causes me concern. Thank 
> goodness 
> it's Friday ;-)
> 
> -- 
> Hugh
> --
> Hugh Taylor
> Head, Collection Development and Description
> Cambridge University Library
> West Road, Cambridge CB3 9DR, England
> 
> email: [log in to unmask]   fax: +44 (0)1223 333160
> phone: +44 (0)1223 333069 (with voicemail) or
> phone: +44 (0)1223 333000 (ask for pager 036)
>