[Koha] MARC Records

[Jonathan Moore]

--------------------------------------------------------
------------------------

>MARC records have an astonishing amount of detail, 
both >in terms of data
...
>possibility of adding a field to bibliosubject to > 
>specify a subject
> heading system used?

The system we are using seems to use the redirection 
entries that someone mentioned. I think we must have a 
whole table of entries that say to place a particular 
marc entry tag into a particular area of the programs 
database.  It also seems to have a similar table for 
exporting.  It would be nice to program this 
flexibility into Koha.  I think it would help down the 
road in supporting other input formats such as the 
differences between USMARC and UKMARC.

I agree with the idea of storing the marc record en 
mass in the database.  There is a lot of use for being 
able to export it out at a later time, plus I think a 
lot of librarians are going to want to be able to 
modify their collection data from a marc oriented 
interface.

>I also have a request.  Could people send me some 
>samples of their MARC
>records so I can do some sanity checking?  A lot of my 
>MARC records are
>not from "formal" sources, but were hand entered by 
>the librarian here.

I have posted a couple of marc record files to our ftp 
server.  I am working on a data conversion to get our 
data out of the system we are currently using to 
something new, possibly Koha.  The old system doesn't 
place everything in the correct place so keep that in 
mind.  Some of the information is being put in the 900+ 
range instead of the "correct" location.  For example 
the one file with bar code information has it in the 
945 range and it should be a sub field of the 852 tag. 
One of the files is quite large 30 megs it contains 
marc records for just about every item in our high 
school collection.  There should be some good examples 
of multiple records that should when processed have a 
single shared marc bib record.

ftp://usd465.com/pub/marc-samples/

I'm also curious to know how librarians typically use 
commercial CDs of
MARC records.  Do they import all of the records into 
their library system
at once, or just import records for books they hold?  
How are the MARC
records typically stored on a CD?  Is it just one big 
file full of MARC
records?

The way our librarians want to use MARC is done by an 
add on to the Follet system.  It is a yearly cd 
subscription of common titles/all titles you are likely 
to have.  There are two main uses

1.  Run it through your existing holdings data and it 
will insert correct mark records for the items.

2. Use it when you get new titles.  The marc cd is 
searchable.  Search using whatever field you desire: 
title, author, isbn, etc.  Once you find the title 
click on add copy and put in copy specific data.  You 
are down.  I think we order books pre bar coded, so I 
think they look up the Marc record, click create copy, 
scan the bar code and enter the price.

Some thoughts on determining when to create a new marc 
bib record in the system.  I think in most cases that 
if you have items with the same title and author then 
they are probably different copies of the same 
intellectual property and hence should be link together 
in the system in some we whether that is by a common 
marc record or some other method.  What if we had 2 
configurable threshholds?   I list of one or more 
boolean conditions that must be true for automatic 
combining and one for queried combining.  For example

If author and title and publisher are all the same that 
the items are automatically combine in some manor or 
cross referenced.  If the author and title are the same 
and nothing else then query the user if he/she wants to 
combine or cross reference. 

What are you using to do the Marc processing?  I have 
been looking at the MARC.pm code with the same idea in 
mind, but I haven't done any coding yet.  I have just 
been looking at the documentation to get an idea of 
what it can do. It seems pretty powerful and flexible. 
The website for MARC.pm is

http://marcpm.sourceforge.net/

PS
I am not sure how this works but I remember reading on 
the Library of Congress site on marc record format that 
there are four kinds of marc records

1. Bibliographic
2. Authority
3. Holdings
4. ???

These could be all wrong I am just going by memory.  I 
didn't see any obvious way that these are flagged 
anywhere.  For example Follet seems to treat any marc 
record without a bar code as a bib record and anything 
with a bar code can end up creating both a bib record 
and a holding record that is tied to it.

Sorry to go on so long, but while we are talking about 
MARC record cds, how hard would it be and what legal 
issues are their on creating a shared marc record 
library for Koha?  Either something that could be 
accessed live from remote, or perhaps a distributed 
network where groups of libraries inter publish marc 
records.  Does this already exist somewhere from a non 
commercial source?


Jonathan Moore
District Technology Coordinator
Winfield Public Schools
316.221.5100 Technology Office
316.221.0508 Office Fax


---------------------------------------------------------------
Visit Winfield Public Schools' website at http://www.usd465.com