On Wed, 29 May 2013 18:58:40 +0000, Tennant,Roy <[log in to unmask]> wrote:

>On 5/29/13 5/29/13 � 9:35 AM, "J. McRee Elrod" <[log in to unmask]> wrote:
>
>>What consistency?  We have in our database "il.", "ill.", and now
>>"illustrations".  To my shock, I found "illus." in a few derived
>>records, which so far as I know was never a standard.
>
>Yeah, well, whatever.
><http://experimental.worldcat.org/marcusage/2013-04-300b.txt>
>Roy
>========================================================================

Yeah, well, whatever.

To make that extract more useful, Roy, may I suggest you first select for 
the language of cataloguing (viz. English, coded eng).  Then sort by the 
cataloguing code/format.

Then it would be quite possible to reduce the apparent wilderness to a few 
patterns, even to create a means of standardizing the vast majority.

Codes have differed in their specifications for such data: it would be more 
useful to know how much variation there is between records created (or 
revised) under the principal codes, and whether machine manipulation of the 
data could reduce the variations to 50%, 90% or 99% without significant loss 
of information.

It might also be useful to see the data over spans of years according to the 
cataloguing codes chiefly in force.

Non-English codes will be a different matter to correlate.  Anything outside 
a standard formula will be harder to handle, but if WorldCat management is 
serious about its international scope, that too will have to be addressed -- 
and coding seems to be the obvious tool.

Until maybe a decade ago, most of us were cataloguing for the benefit of the 
human eye (of the reader); now we're expected to catalogue for the machine 
(while still yielding results interpretable by the human eye).  It's hardly 
fair (as you seemed to imply, but maybe I'm wrong there) to complain that 
results of our past work now don't suit the machine and the algorithms it 
applies.

Hal Cain
Melbourne, Australia
[log in to unmask]