Re: Filing-rules sort subroutine for authors' names?

2004-07-26 Thread Ben Ostrowsky
Andy Houghton wrote:
Don't confuse ALA filing rules with NACO normalization rules.
If you are trying to compare two author names you should use
the NACO normalization rules [1].  If you are trying to sort
the headings into order then use the filing rules [2].
We're trying to get a shelf list, and when the call number is "F LAM", we 
can't just sort by call number -- and when the authors' names may be given 
as "Lambert, John", "La Marche, Susan", and "Lamarque, Erich Maria", we 
can't just sort them ASCIIbetically either.

So I guess it's the filing rules I'm really after.
Ben


RE: Filing-rules sort subroutine for authors' names?

2004-07-26 Thread Houghton,Andrew


> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] 
> Sent: 26 July, 2004 13:58
> Subject: Re: Filing-rules sort subroutine for authors' names?
> 
> Definitely possible--library automation systems and card 
> printing systems do it.
> 
> I'm not fully conversant with the rules myself, but can tell 
> you that it's a lot more work than a regex plus a string 
> compare.  (I'm thinking of sorting titles here, actually.)  
> Handling diacritics, sorting "10" after "9", etc. adds up to 
> a decent amount of work.  It's worth your while to determine 
> whether you have to implement the rules 100% or have some wiggle room.
> 
>>From: Ben Ostrowsky <[EMAIL PROTECTED]
>>Subject: Filing-rules sort subroutine for authors' names?

>>
>> Just a sanity check: is it really possible to create a Perl 
>> subroutine that
>> 
>> would compare two authors being sorted and enforce the ALA 
>> filing rules?  I
>> 

Don't confuse ALA filing rules with NACO normalization rules.
If you are trying to compare two author names you should use
the NACO normalization rules [1].  If you are trying to sort
the headings into order then use the filing rules [2].  You
can probably write a simple Perl routine to do the NACO 
normalization rules.  The filing rules are much more complex,
since you need to take into account numbers, e.g. 9 vs. 999,
as well as Roman numerals, dates in a variety of formats, 
including spans, and articles in various foreign languages.
So it's a lot more complex to do ALA filing rules correctly.

IMHO, it's almost impossible to do ALA filing by computer, due
to the rules, and having tried this for several concordances 
between LCSH and Dewey.  My last attempt was for the publication
People, Places & Things [3] where the Editors only found four
headings that were miss-filed out of 60,000.  And this was a
small subset of LCSH.  With a lot of work you can get it mostly
correct...


Andy.

[1] http://www.loc.gov/catdir/pcc/naco/normrule.html
[2] "ALA Filing Rules", Americal Library Association (ALA),
(c) 1980, ISBN: 0-8389-3255-X
[3] "People, Places & Things", OCLC Online Computer Library
Center, Inc., (c) 2001, ISBN: 0-910608-69-5


Andrew Houghton, OCLC Online Computer Library Center, Inc.
http://www.oclc.org/about/
http://www.oclc.org/research/staff/houghton.htm



Re: Filing-rules sort subroutine for authors' names?

2004-07-26 Thread moregan




Definitely possible--library automation systems and card printing systems
do it.

I'm not fully conversant with the rules myself, but can tell you that it's
a lot more work than a regex plus a string compare.  (I'm thinking of
sorting titles here, actually.)  Handling diacritics, sorting "10" after
"9", etc. adds up to a decent amount of work.  It's worth your while to
determine whether you have to implement the rules 100% or have some wiggle
room.

Mike O'Regan
Follett Library Resources



   
   
  Ben Ostrowsky
   
  <[EMAIL PROTECTED]To:   [EMAIL PROTECTED]

  >cc: 
   
   Subject:  Filing-rules sort subroutine 
for authors' names? 
  07/26/2004 12:43 
   
  PM   
   
   
   
   
   




Hey guys --

Just a sanity check: is it really possible to create a Perl subroutine that

would compare two authors being sorted and enforce the ALA filing rules?  I

haven't read 'em, but I'd imagine that some sort of real-world knowledge
might be called for.

Ben