BibIndex: new tag for address field

[fpoli]

• Associates the address field to a single 371__% tag instead of many
  371__a, 371__b, 371__c, ... tags
• Keeps other tags associated with the address field: 110__a, 110__b,
  110__t, 110__u, 410__g

Signed-off-by: Federico Poli federico.poli@cern.ch

[jalavik]

@fpoli To confirm the actions needed to be taken on production:

change address field tags from:
110__a, 110__b, 110__t, 371__a, 371__b, 371__c, 110__u, 371__e, 371__f, 410__g
to:
371__%, 110__%, 410__g, 410__a

@kaplun I guess then we should re-index address fields?

[jalavik]

Rebased as: 02ab63d

[kaplun]

@fpoli where is the specification of this change?

@jalavik yep, we should reindex the whole address index.

[aw-bib]

Please note a slight difficulty here: 371is valid only in context of *authority* records, while110and410` are valid in bibliographic as well. However, there is a slight different semantics between those in authority and bibliographic.

• In authority context 110 and 410 refers to the main form of an organizational name (ie. records that describe an institution like CERN), 410 to an alias for 110. E.g. 1101_ $aEuropean Organization for Nuclear Research with 4101_ $aCERN.
• In bibliographic records 110 is name of the issuing institution, while 410 refers to a series statement e.g. in book series. E.g. Annual report of 1101_$aCERN, published in 410__$aCERN Yellow reports.

[fpoli]

The (quick) specification on Asana: https://app.asana.com/0/11661778102949/12871817123950

[jalavik]

Assigning this to @kaplun

[kaplun]

@aw-bib we are discussing here how to best tune bibindex so that users searching for addresses will actually find records. So in general the more the better, unless the given field is used for a given algorithm, but in this case we tend to use more specific definition such as even directly using the MARC.