[OPEN-ILS-GENERAL] Authority Control Processing

[Mike Rylander]

On Thu, May 15, 2014 at 9:37 AM, Martha Driscoll <driscoll at noblenet.org> wrote:
> The discussion of authority control vendors reminds me that I don't have a
> solid grasp of authority record processing.
>
> This is what I have been able to figure out.  Let me know if I have made any
> incorrect assumptions.
>
> Let's assume a 2.5 system.
>
> Initial record loading
>
> 1. Authority records are loaded into the system.
>
> 2. The script authority_authority_linker.pl is run.  This script links
> reference headings in authority records to main entry headings in other
> authority records.
>
> 3. The script authority_control_fields.pl is run.  This script goes though
> each bib record and creates a $0 in any field with a matching authority
> record.
>
> Does it matter what order this is done in?
>

I would strongly recommend authority_authority_linker.pl first, to
avoid causing yet more (unnecessary) bib reingest.  Also, there is a
bug fix for authority_authority_linker.pl at
https://bugs.launchpad.net/evergreen/+bug/1312945 that you'll want.

> Ongoing Processing
>
> 1. As new bibs are added to the database, run authority_control_fields.pl on
> the range of new bib id's.
>
> 2. As new authority records are added to the database, run
> authority_authority_linker.pl on the range of new authority record id's.
>

Sounds good.

> Do old bib records ever need to be run through authority_control_fields.pl
> again?

Not unless they are edited, which brings us to ...

>  What if they are updated?

The right-click authority mechanism can be used for controllable
fields during human edits.

>  What if they are overlaid via vandelay?

Probably good to re-link these.

> Is there ever a circumstance where a new or updated authority record now
> matches old bibs?

Sure, that could happen ...

>  How would you identify the bibs, or do you have to run
> authority_control_fields.pl through the bib database periodically?
>

Actually, the browse data (metabib.browse_entry, and the bib- and
auth-related linking tables that point at it) could probably identify
most of this.  There's no preexisting SQL for that today, but all the
data is theoretically there, and the bib browse interface effectively
uses the same logic you'd need to find unlinked-but-matching headings
in order to fold all headings together.

> --
> Martha Driscoll
> Systems Manager
> North of Boston Library Exchange
> Danvers, Massachusetts
> www.noblenet.org



-- 
Mike Rylander
 | Director of Research and Development
 | Equinox Software, Inc. / Your Library's Guide to Open Source
 | phone:  1-877-OPEN-ILS (673-6457)
 | email:  miker at esilibrary.com
 | web:  http://www.esilibrary.com