Just yesterday I was attempting to process a MARC file from the National
Library of Scotland (I *think* they use UKMARC, which is sufficiently
close to MARC21 that I don't need to know :). The processing stopped
on record 9190 (of 113296) because of an alpha indicator (admittedly it
was an "o", and I suspect it was meant to be a 0, but I'm not sure
about that).
I really think that this change, which involves typing in three more
characters into the code, is justified simply because it's part of the
specification. It's not some wacky idea someone's had that would be a
neat thing to do with a MARC record. Surely it is much better for the
code to cope with the spec when someone needs it, rather than that
person needing it first and having to wait for the development cycle?
I know I would have much rather my files processed properly yesterday
without me having to add to this thread!
Don't get me wrong though, we're very appreciative of all the work that
has gone into the MARC::Record suite -- it is *extremely* useful
[there's an implicit thanks in there]!
Cheers,
Ben
On Wednesday, 19 November, 2003 19:55, Morbus Iff wrote:
> >Yup, looks like it. This is a new one on me, who haven't really
> > studied MARC since I escaped from MARBI in 1992. Maybe this is a
> > USMARC->MARC 21 change, i.e. something that came in with MARC 21?
> > Or was it a change in the ANSI or ISO standards? I don't think
> > lowercase alphabetics were valid indicators in USMARC, which is
> > why I'm positing a change somewhere, although my memory may be
> > deficient. In any case, the LOC statement as to current standard
> > seems quite clear.
>
> As Ed said, and I'm slowly starting to agree with for different
> reasons, does anyone absolutely need the alphabetics as an indicator?
> More importantly:
>
> * what should MARC::Record read and write as a baseline?
> MARC? USMARC/MARC21? UniMARC? I don't even know the
> differences between them, but is MARC21 backwards
> compatible with MARC? Is anyone who's anyone ALWAYS
> using the latest LC MARC standard? Is that the same
> assumption that should be placed on MARC::Record?
>
> * If someone is ONLY using MARC, and they MARC::Lint,
> will they expect MARC-rules ("no alpha indicators") or
> MARC21 rules ("alpha indicators"). How badly will this
> break Lint scripts? A MARC record shouldn't be validated
> against MARC21 rules (especially if another program
> expects non-alpha's under MARCs, but gets a linted and
> "correct" MARC21 with alphas).
>
> Note to readers: When I say "MARC", I mean "the one that no one
> remembers allowing alphabetic indicators". When I say "MARC21", I
> collectively mean "USMARC/MARC21" or "the one Anne quoted that says
> alphas are dandy". I suspect I'm using one of those terms incorrectly
> - cluestick happily.
--
Ben Soares tel: +44 (0)131-651 1238
EDINA, Edinburgh University Data Library fax: +44 (0)131-650 3308
Main Library Building, George Square email: [EMAIL PROTECTED]
Edinburgh EH8 9LJ, Scotland, UK www: http://edina.ac.uk/
"Hmmm, that makes no sense to me...
But then you are very small, perhaps you're right." -- Treebeard
