Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 6 Next »

INSERT LINK TO MSWG DOCUMENTATION WHEN READY for prospective consideration

BACKGROUND

Upon migration to Alma, Harvard will be using Alma's automated authority control process. This process attempt to link every access point to an LC authority record. The Alma Resource Management Working Group has done significant testing of this feature and found that it works well about 98% of the time. This feature will be extremely helpful in keeping our headings up to date. For example, we currently use MARS to inform us when a death date has been added to an author in the LC record. We then have to go into Aleph and correct hundreds of headings either manually, or in small batches. With Alma, if an authority changes, all our corresponding bibliographic headings will automatically change. 

The system works through a process where the bib headings are linked to a corresponding authority record. (Similar to OCLC when you control headings.) However, in OCLC a cataloger has to manually control any heading that is unqualified. That is, a heading with $a but no $b, $c, $d, or $q. 

In contrast to OCLC, in Alma the headings, even if they are unqualified, will be automatically linked to an authority record, as long as the heading exactly matches a 1XX or 4XX on a single  authority record. The matching (or controlling in OCLC terms) happens when a record gets saved to the server. Then during the night, the headings that matched a 4XX in the LC record gets flipped to the 1XX field. Therefore, there is the possibility of headings being incorrectly flipped.

For example, a book on cats has the author, Quintana, I. There is this authority record:

  • 100  Quintana, Isabel
  • 400  Quintana, I.
  • 670  I only ever write about medicine, 2018

There is an NAR that will match. However this NAR is for a doctor, who is NOT the same person as the person who writes about cats.

 As you can surmise, if the headings are qualified (by $c, $d, etc.) the chances that they match to the correct NAR are high. However, Aleph has 4.5 million unqualified headings. We tested a set of around 7000 headings, and we found that only 2% matched a 4XX in the NAF, and only .02% matched a different entity. So the grand majority of headings will match the correct entity.

However, we are still trying to be cautious. Some folks who work in Special collections have asked us to exclude their retrospective records from this process. The way we hope to do this is to add $c (Harvard local name) to the headings on these records, so that the headings will be prevented from matching a record in the NAF.

Since I know that there are few surnames in CJK languages, we wondered if we might also want to exclude HYL records from this process. Likewise we wanted to alert Judaica and Middle East, in case there were certain issues at stake for these languages.

SUMMARY

For our retrospective data:

  • You can opt to have all bib. records where HYL, or Judaica or Middle East is a holdings treated so that ANY unqualified headings will have $c (Harvard local name) added to them. Please note that this means that even if the name is correctly controlled to an NAR in OCLC, the same 1XX/7XX in Alma will have $c (Harvard local name) added to it.
  • You can choose to leave your data as it is. The headings will then go through the Alma authority control process. The resulting reports will be too long for us to review them, as we will have 4.5 million unqualified headings being matched to NARs. However, from our sample only .02% were linking incorrectly, so most headings either don’t match at all, or match correctly to an NAR.

For our prospective data:

  • You should be careful whenever you add an unqualified name to a bib. record. MSWG will be issuing some guidance in this area soon. Any unqualified name will run the risk of being linked to a future NAR that may not necessarily represent the entity on your bib. record. You can always choose to add a qualifier to your names to make them more unique, if you think this is a problem.
  • Prospectively you will be able to scan the authority matching output report, so you will be able to check if headings matched the correct entities, and add a qualifier to disassociate your name from the authority record, if necessary. 

SCOPE OF WORK

  • Fields included 100, 600, 700
  • Definition of unqualified name heading: contains a but not any of: b,c,d,q.
  • Excluded: PCC/DLC records
  • Included unqualified names for these categories: 
    • Heading included $5
    • Heading included $e with one of these values (case-insensitive)(list compiled based on record counts): 
      • Former owner
      • Printer 
      • Ill. 
      • Ed. 
      • Collector 
      • Tr. 
      • Engraver
    • Library holdings for: HYL, HOU, POE, THE
    • Language of bib: Hebrew, Yiddish, Arabic, Persian, Kurdish

 

FUTURE

can remove these from sets of alma records at any time after migration

refer to new MSWG policy re: qualifing headings when they conflict with NAR

 

 

  • No labels