Transcript Document

FAST AT CORNELL
FAST Interest Group, ALA Annual Conference 2014
Steven Folsom, [email protected]
Chew Chiat Naun, [email protected]
Collaborating
with OCLC
Research for
Enhanced
Discovery
THE FAST TEAM
OCLC Research
Rick Bennett
Eric Childress
Kerre Kammerer
Ed O’Neil
Diane Vizine-Goetz
Cornell University
Gary Branch
Chew Chiat Naun
Steven Folsom
Sarah Ross
Ardeen White
Cornell catalogers: Yen Bui, Roswitha Clark, Sung Ok Kim,
Yelena Kurbanova, Apikanya McCarty, Teresa Mei
Cornell Discovery and Access Team: (Too many to list)
BLACKLIGHT CATALOG
Open Source
 Greater control over Indexing Decisions
 Greater control over User Experience
Fits within our larger Discovery and Access
Initiative
 Co-indexing local digital collection metadata
 Co-indexing third party content, e.g. Hathitrust DOAB
CORNELL PIPELINE FOR
BLACKLIGHT
•MARC21
(from ILS and
elsewhere)
•Non-MARC
Data
Sources
Integration
Layer
•XML
•RDF
•Mapping to
Solr Fields for
Blacklight
Indexing
Q: WHY THE TRIPLE STORE?
A: Inferences
FOR EXAMPLE…
ULAN TO LCNAF
LCNAF TO FAST
CURRENT SUBJECT MAPPING IN SOLR
CURRENT SUBJECT FACETING IN BLACKLIGHT
SUBJECT/GENRE
PARTNERING WITH OCLC
December
11-12, 2013ALCTS
eforum, OCLC
offers to
partner with a
library to
convert their
LCSH. Cornell
requests
service.
September
9th, 2013Cornell
contacts
OCLC about
2001
publication
September
2013- OCLC
Announces
project to
enhance
WorldCat db
with FAST
headings
Spring 2014OCLC and
Cornell
discuss
feedback on
problematic
mappings
Spring 2014OCLC and
Cornell test
Batch
Conversions
of Cornell
Records
Near FutureMerge FAST
headings into
Cornell MARC
records
Spring 2014OCLC and
Cornell
strategize
methods for
extending
FAST
CATALOGING WITH FAST
 Pilot began February 2014




7 catalogers currently participating
For minimal-level cataloging only (25% of our original output)
Replaces use of uncontrolled 653
Currently using http://fast.oclc.org/searchfast/
SEARCHFAST
SEARCHFAST
CATALOGING WITH FAST
 FAST vs LCSH
 Introductory presentation
 Characteristics of FAST
 Strengths and weaknesses
 Cheat sheet
 FAST “mindset”
KEY TRAINING POINTS




Use what you find (all headings established)
Subjects do not cross facets
Observe distinction between topical and genre/form facets
Fewer application rules
• No constraints on combinations of topical and geographical terms
• Order of headings is not significant
 Dates can be whatever you need to assign.
DATES
VOCABULARY ISSUES





Unestablished headings
Problematical mappings
Missing references (e.g. vernacular)
Scope notes
Event headings
WHAT HAPPENS IF A HEADING IS NOT IN
FAST?
Extensions
 Not in LCNAF:
600 17 |a Folsom, Steven |2 fast/NIC
 In LCNAF but not in FAST:
600 17 |a Childress, Eric |a fast/ naf |0 (DLC)no2005043559
 Legal but not explicitly established in LCSH:
650 _7 |a Elephants |x Weight |2 fast/lcsh
Headings in our conversion project that failed to
validate against FAST will be given similar treatment.
PROBLEMATICAL MAPPINGS
“RELATIONS” – OCLC SOLUTION
Relations [countries]  International relations
Relations [religions]  Interfaith religions
Foreign relations  Diplomatic relations
Foreign economic relations  International
economic relations
Military relations  Military relations
EVENT HEADINGS
FAST conversion output
611 7 $a Trenton, Battle of (New Jer sey : 1776) $2 fast
648 7 $a 1776 $2 fast
651 7 $a New Jer sey $z Trenton $2 fast
FAST event authority
034 __ $d W0744435$e W0744435$f N0401301$g N0401301$2
geonames
043 __ $a n-us-nj
046 __ $s1776
053 _0 $a E241 .T7
111 0_ $a Trenton, Battle of (New Jer sey : 1776)
511__ $a American Revolution (1775 -1783)$0 ( OCoLC)fst01351668$w g
551__ $a New Jer sey$z Trenton$0 ( OCoLC)fst01 207908
STATE OF PLAY: CATALOGING




About 450 records in first 3 months
Unit time about 50% higher than minimal-level cataloging
Well-accepted by catalogers
Further steps
 Maintenance
 Improvements to editing environment (maybe)
 The Elephant in the Room
FINALIZING BATCH PROCESSING AND MERGE
 Add valid FAST headings and local extensions to
records
 For names not in LCNAF, add $2 fast/NIC
 Recode $x subdivisions to $v if they match a current
$v subdivision, in LC cataloged records created pre 2000 and all non-LC cataloged records. (Hoping this
to be a little more nuanced.)
 Retain original LCSH headings in the records
TENTATIVE SOLR MAPPING
Vocabularies for faceting
 Most likely to only use |2 fast, |2 fast/NIC, and |2
fast/naf
 Could conditionally use $2 naf if there are no FAST
headings since the literals would converge
Conditional logic for item record display
 Likely to supress FAST from item view if LC
authorities are available.
PROJECTING PROGRESS: FACETS
After
Before (Now)
S u b je ct / G en re
G a m blin g
Ca s inos
M i n o rit y bus i n ess e n te rpri s es
Un i te d St a te s . Sm a l l B us i n ess
Adm i n ist ra t ion
B i l dun gsro mans
12
4
2
2
1
S u b je ct / G en re
G a m blin g
C a s inos
M i n o rit y bus i n ess e n te rpri s es
U n i te d St a te s . Sm a l l B us i n ess
Adm i n ist ra t ion
B i l dun gsro mans
Se e m o re > >
S u b je ct : Re g io n
La s Ve g as ( N ev. )
17
N eva da La s Ve g as
11
Un i te d St a te s
2
Ari z o na B l a c k M e s a ( N ava jo Co un t y a n d
Apa c h e Co un t y )
1
B l a c k M e s a ( N ava jo Co un t y a n d Apa c h e
Co un t y, Ari z . )
1
Se e m o re > >
S u b je ct E r a
2 0 th c e n t ur y
4
12
4
2
2
1
Se e m o re > >
S u b je ct : Re g io n
N eva da - - Las Ve g a s
U n i te d St a te s
A ri z o na - - Blac k M e s a ( N ava jo Co un t y
A pa c h e C o un t y )
28
2
and
2
Se e m o re > >
S u b je ct E r a
1 97 0 - 1979
1 97 8
2 0 th c e n t ur y
3
1
1
PROJECTING PROGRESS: ITEM VIEW
Before (Now)
After
OUTSTANDING ISSUES
Assessment
 Workflow… Is it fast enough?
 Do mappings need tweaking?
 Is genre data good enough to parse out from topical?
Digital collection use of FAST
 Legacy metadata… Convert or infer?
 New Digital Collections… use FAST as vocabulary or
map?
MAINTENANCE
Still working out the details, but:
 OCLC will supply report of changed headings
 IDs should greatly facilitate automation
 OCLC master record updates
Not restricted to ILS data
FAST CHANGE REPORT
RESOURCES
 O’Neill and Chan’s FAST: Faceted Application of Subject Terminology
(OCLC #624025531)
 OCLC FAST web site:
http://oclc.org/research/activities/fast.html
 ALCTS eforum summary:
http://lists.ala.org/wws/arc/alcts-eforum/2013-12/msg00131.html
 Cornell cheat sheet: http://lts.library.cornell.edu/lts/pp/cat/127FAST
 Cornell FAST introductory presentation:
http://ecommons.cornell.edu/bitstream/1813/34435/2/FAST_C%26M.pptx
 Other local Cornell documents available (tips, problems, etc.)
QUESTIONS?