Notes
Shelby E.
Harken
ALA Annual –
San Francisco – June 14-19, 2001
8:30-5:30 Hilton & Towers -- Continental Ballroom 5
Future of Serials Control : implementation of the
MARC21 Holdings Format
See conference
packet
Sally Sinn – chair
of the CONSER Pattern Committee
Sally feels we are on the verge of making real use of the format in a way that will reduce the redundancy of everyone having to figure out checkin patterns on their own.
Frieda Rosenberg
Holdings display standard is Z39.71,
a NISO standard
Holdings communications standard is
MARC21 Format for Holdings Data
Three related standards: Z39.42,
Z39.44, Z39.71
Why should we use
standards?
They make records transferable / we
can migrate records more easily
It is easier to enhance records
We can buy records
We can do resource sharing
Some issues:
What type of interface is necessary – tagged or GUI? – we always need both, with access to tagged required
A
linked record is better than embedded
Whether items are dependent or
independent affects display of individual and/or summarized
Coded
holdings or free text (which are harder to deal with)
Single
copy per record (works better) vs. multiple
Multiple versions or not linked to one
bib record (possible but sometimes results in unclear displays
Summary holdings can
be put in the textual holdings field 866
Single
item holdings may be in one 853 field
Piece holdings (i.e. more than one item
per bib record), use paired 853-5 with 863-5 with #8 linking fields
Holdings
records fields:
Type of record: v = multipart item holdings, x = single part item holdings, y = serial item holdings
Elvl: 3, 4, 5 is to say HOW detailed your holdings are, not for listing your summary holdings. If all you do is summary, use #3. If you use 87X item tags, then Elvl is 5
Receipt
through Retention are for ILL and union listing
Language, use the MARC code; the intention is that the holdings will cause display of holdings in that language.
007 would be same as bib except if you are doing mulver
Notes: 541 – source of acquisitions, 561 – custodial history, 583 – action note with 19 possible subfields; 842 – textual physical form in natural language; 843 – reproduction note; 844 – name of unit; 845 – terms governing use and reproduction
852 #a UND often doesn’t display (would you in a consortia?) #b location #c shelving #h class no. #I cutter number #x note #z public note
852 indicators
1 = shelving scheme 0 (LC), 2 (DDC). 3 (SUDOC). 5 (by title), 6 (class sep)
2 = shelving order 0 = classed separately, 2 shelved by secondary enumeration
853 Captions and Pattern field, 854 for suppl. And 855 for indexes
#8 number will match with 863’s #b pattern – this is what triggers prediction
The use of parenthesis keeps words from displaying, e.g. #j (month)
#u = how many pieces you get before vols. roll over?
#v = do the numbers restart or continue with a new year?
#w,
x, y – apply to the serials as a whole
863
second indicator
0 = range, 1 = itemized, 2 = we want to display textual holdings in 866-8, 3 = one unit, display from 866 text, 4 = items not published
Look
for systems that:
Display
the order you want
You
must be able to display and edit tagged records
You must be able to re-order fields, so you don’t have to renumber all the #8’s when you get an old issue
VSTF!
EpixTech does NOT use MARC21 Holdings Format!
Use
questions in survey in workshop handout to ask vendors
See
handout, p. 9, #50-51
EOSi
review
(see
booklet – Karen Anspach)
Can search serials (which are “automatically associated” with bib records) by a variety of search keys
There is an enumeration file the
library can create and put in drop down menus
Fully supports actions with all
subfields
852 is used for each copy, useful if
separate copies are in different locations
86X tags are automatically created at
checkin and created 876 with barcode (very little in it)
Edit button can be used on any checkin
window
ExLibris (Helen
Gbala)
Next summer Aleph will be 100%
MARC21 compliant
Can use multiple 853’s using #8 to show freq change historical and when to begin new pattern
Plus sign is used to show ordinal
number
Enumeration and chronology supported
except for X which is completely irregular; use #y
Currently a few fields are not
supported and codes for “day” got reversed; they will fix
Batch load to pull 891’s from OCLC –
they are working on, especially for migrating libraries
Holdings record links to bib and is
control for all items; must have a bib, then holdings, then administrative
record, create an order record “subscription record”
Publication schedule is created from
holdings and subscription record
Checkin possible via SIC, in
publications window, or in list of items window
An item record is created as part of
checkin and assigns a barcode if you want it to
Repeatability of #y and new #z are
likely to not be working yet in a number of ILS systems – Endeavor (not yet),
ExLibris (next summer), EOSi (not yet)
Current serials
holdings standards – Rebecca Guenther
CONSERHOLD is a listserv that
discusses changes in holdings format
MARBI DP 10 #y only addresses
chronology, not enumeration at this time
How should we handle e-holdings?
856 – put general site
in bib
856 – put local license
in holdings
Serials Pattern
Project
See http://www.loc.gov/acq/conser/patthold.html
Hope to keep 891’s updated
If the frequency changes, there will be
paired 891’s with new 853 and 863 in it
Pilot project for Serials
Pattern Project
To do a 2-year pilot – to get
853/863 in serial bib records
Analyze workflow and make
recommendations
Involve vendors
Recommendations to improve MFHD
Do documentation
Consider roles of utilities and
subscription vendors
Publicity
Doing current serials only
(eventually old titles will have because titles change)
Dealing with a standard
Voyager checkin has underlying assumptions that
result in a number of problems:
·
Patterns operate
essentially at templates – not optional nor can you use them to build on for
slightly different situations
·
Numbers of patterns
are finite – NOT !!
·
Users don’t want to
deal with coded data – some of us do!!!
·
User-friendly
front-end saves time – sometimes it doesn’t!!
·
Simplicity is
desirable and possible with serials
·
Prediction the
primary functions of checkin systems – NO!!
·
Limitations of
template strategy
o
It aggregates work
o
It is incapable of
handling variant languages to display captions (have to have different pattern
if you want to display eg. Norwegian, yet there is a language code in MARC21
that SHOULD CONVERT the language automatically)
o
Dependent on
proprietary algorithms (i.e. Voyager programming), rather than MARC21
o
Snowball effect –
once you have a bad form, you either end up with too much editing or
increasingly bad records
o
What makes a unique
pattern?
§
Order and
arrangement of enumeration and chronology
§
Language
§
When calendar
changes
§
Issues
combined/omitted
o
Standard data cannot
be imported or exported (i.e. can’t use 891 in OCLC records) – this affects
both union listing (MULS) and the Pattern Initiative to share checkin patterns
via OCLC records
o
Current
implementation is rigid and inflexible
o
Endeavor’s plans for
enhancements is insufficient to support sharing of patterns within the
foreseeable future
·
Functional problem
areas:
o
Combine OR omit are
your only choices – you can’t do both
o
Limited languages –
only 6 or 7 are supported
o
Cannot handle
ordinals – we have these!
o
Cannot handle
captions after enumeration – we have these!
o
Continuous numbering
at second level is problematic
o
Collapse function is
DYSFUNCTIONAL!
·
Snowball effect
o
When you get to the
end of the year and then don’t get a Thanksgiving or Christmas issue, it thinks
you didn’t checkin “correctly” so the pattern regenerates wrong, so you have to
close the pattern; then you restart at the beginning of the year. If you do
that, then you can’t claim on last year. Voyager can predict 50 issues, just
not the right ones.
·
Voyager libraries,
can
o
Use the Bremer macro
to create patterns
o
Download to Voyager
o
Copy/paste holding
to record, code to differentiate from a collapsed holding by Voyager dates – or
collapse and edit – but if you collapse, it freezes checkin
·
Voyager has an
enhancement committee trying to identify problems which boil down to lack of
compliance with MARC standards
VTLS
(John Espley) moves only the 85X from the 891 fields in records, not the 86X
(which is more than Voyager). He gave a very “straight talk” presentation
describing what they do using MARC standards
Panel (Berit Nelson, DRA; Kathryn Harnish,
Endeavor; Karen Anspach, EOSi; Helen Gbala, ExLibris; Ted Fons, III; John
Espley, VTLS)
·
DRA – uses 876-8 if
you wish to use MARC for items for circulating. Linking order and sequencing is
hard to derive; it’s easier to have separate holdings records for each copy
·
Voyager – only
brings in 891 to 853/4/5 – this is contrary to Diane Hillman’s presentation
(trusted source); “glossy” presentation
·
III – form approach
to summary holdings – like the OLD 590 approach on bib records; they have
checkin patterns for new holdings and different patterns and methods for
historical holdings – why? He didn’t think they can do 876-8 – i.e. NO ITEM
LEVEL HOLDINGS!
·
891 – how do they
each use?
o
VTLS already creates
holdings from 891
o
ExLibris does and
they match on ISSN; looking at pulling OCLC bibs as a source for getting back
holdings for customers (forward thinking idea)
o
DRA – no specific
loader; it is ‘easy’ for libraries with no holdings because there are no
overlaying issues (does that mean you are out of luck if you have holdings???)
o
Endeavor – not now;
looking to the future
o
EOS – share files
with other customers (probably not standards-based)
·
What if we just want
891 and not the rest of the record?
o
ExLibris – partly
dependent on how OCLC can deliver (honest answer); right now can bring in
record, pull data from 891 to create holding, then delete
o
III – generates an
entirely new bib record or matches on match points, then only generates new
holdings record
o
VTLS –
record-by-record; can cut-and-paste
·
Validation
capabilities
o
VTLS can validate
any MARC format either before or after save
o
DRA – hmmm
o
ExLibris validates
what you want to validate and at what level with appropriate message; full
validation of all MARC formats – I think this means you have control of
validation tables and perhaps message, but it validates when you are ready to
submit the record (MARC21 standard available online, as are authorities, and
help is readily available)
o
EOS similar
o
III – incoming
validation only – never later during editing???
o
Voyager – checks tag
tables as you save to the
database
·
Bound-with and
analytics – we want holdings to display to user on BOTH main and analytics
records – note: 004 is non-repeatable
o
DRA handles it in th
item level record but location code must be the same as in the circ location
o
ExLibris – uses a
non-MARC linker field
o
EOS, III, VTLS –
same
o
Voyager uses item
·
Training
o
ExLibris – has
person on Z39.71 committee; MARC21 holdings is discussed in training for
cataloging, serials, and standards (the most straight-forward and thorough
answer)
o
III – it’s simple,
you don’t need training (REALLY!?) ; when you’re ready for advanced training on
the format you can get it
o
Endeavor – dependent
on library needs; how do they learn about MARC, etc.? – they have librarian
developers
o
EOS – train on GUI;
only train MARC if library asked (I expect it in the initial contract!); they
learn from users;
o
VTLS – part of
standard training to use MARC21
·
Use of MARC for
electronic resources?
o
VTLS – focuses on
what’s available to your users; use 852 if you own it; 856 has your link
o
ExLibris – same but can
handle 856 in bib or in holdings which works better in consortia
o
III – looking at
checkin of electronic documents; they are using an XML server to take in issue
level information
o
EOS – issue-level
hot-linking
·
What about something
like JSTOR for back issues and another aggregator for current issues?
o
ExLibris – uses
multiple holdings records
o
III – holdings
record or even item-level
o
VTLS – summary
o
Endeavor – would
depend on how I order the item (what if I don’t “order” it?? – so much in
Voyager hangs on purchase order)
Summing up
7:30-9:30
PM Argent Hotel – Franciscan
III
http://ublib.buffalo.edu/libraries/units/cts/olac/capc/
Authority tools have
been posted at their website
http://ublib.buffalo.edu/libraries/units/cts/olac/capc/authtools.html
Source of title for
Internet resources report is on the CAPC website. The OLAC wqebsite has links
but some links in examples don’t work so main report has web page images
captured and stored
http://ublib.buffalo.edu/libraries/units/cts/olac/capc/stnir.html
Chapter 9 – Should
be published by the end of July, no idea in what form; plus the appendix will
be available; someone is working on updating the index
Abstract writing
primer – this will be on the website; it includes bibliography and examples,
i.e. 520
GMD survey – Jean Wiehs’ report from survey: http://www.modpublishing.com/Survey/GMDSurvey.htm
(Report will be on
CAPC web page)
156
responses with 131 survey’s filed out; remainder either didn’t answer or just
sent a letter
GMD’s
not understood by anyone: realia and kit
GMD’s
users claimed to know what the words meant better than catalogers
Wide
variety of responses to the questions about GMD for music CD’s, or CD’s with
music and video, yet people thought overwhelmingly that GMD’s are critical
Videos
and sound recordings are the two most common GMD’s
Issues
for GMD’s:
o
Preservation of
standards for cataloging – it costs to change; there seems to be a
misunderstanding of the role of the GMD
o
General vs.
specific:
o
If you use general,
you must read more of the record; there is concern about users of web catalogs,
especially when unmediated
o
If you use specific,
there are problems getting international agreement, consistency in assignment,
and recently there has been rapid development of technological types of media
o
Media definitions in
AACR2
o
Activity cards,
games, toys, realia
o
Arti originals, art
reproductions, charts, flash cards, pictures, posters, study prints
o
Electronic resource
o
People are deeply
unhappy with “computer file”
o
“electronic” is
confused with equipment
o
discarded format,
e.g. filmstrips, lose meaning
o
would prefer
separate GMD’s for local and remote electronic resources
GMD’s
– some libraries have indexed GMD’s so they can be used as a limiter
SAA will be publishing AMIA compendiam for moving image collections; some info will be on their website
9:30-12:30 Hyatt Regency –
Seacliff C/D
MARBI –
260 repeatable, MARC21 to accommodate seriality, MARC holdings
See Proposals 2001-04 through 2001-11
and Discussion Papers 2001-04 through 2001-11 at http://www.loc.gov/marc/marbi/an2001_age.html
http://lcweb.loc.gov/marc/marbi/list-p.html
http://lcweb.loc.gov/marc/marbi/list-dp.html
Proposal 5: Seriality – passed with
changes – new BLVL “i” for integrating resources
Loose-leafs and serials with multiple frequencies – MARC Holdings Format records can be repeated for the bib record with related checkin prediction because each holding record can have its own frequency
The proposal indicated it referred to
Type:a but really you still pick the right Type and add 006 for seriality
They decided to leave varying frequency
alone and look to 853 etc. for handling specifics
Proposal 4 – passed
Decided #3 is possible
for all 260’s
Latest is used for
integrating resources with notes about earlier
Earliest is used for other serials with
subsequent 260’s listed in 260’s with indicators
844 should be able to handle
standing order titles without a number designation
Discussion papers were generally
well received
2:00-4:00 Hilton and Towers — Continental
Ballroom 4
SAC
Metadata Program - SAC
Subcommittee on Metadata and Subject Analysis
I chaired the program – we had
nearly 350 in attendance and good questions at the end
Final report, on
Diane’s page: http://www.govst.edu/users/gddcasey/sac/msafinalreport.html
Final
presentations will be on ALCTS web site -- ADD LATER
PERHAPS
AT: http://www.ala.org/alcts/organization/ccs/sac/subjecta.html
Subject Access and
Classification in Metadata for Digital Resources
What kinds of
subject and classification tools do we need to enhance and refine resource
discovery of Internet resources? Speakers will address the Subcommittee's
report, Subject Data in the Metadata Record. Its recommendations can be
employed to find standardized or automated means to efficiently create subject
metadata. Included will be: a comparison of various metadata schemes' subject
analysis, brief explanation of what FAST is, and actual library experience
employing subject analysis.
SUBJECT ANALYSIS AND CLASSIFICATION
IN
METADATA FOR DIGITAL RESOURCES
Program
SAC
Subcommittee on Metadata and Subject Analysis
Diane
Dates-Casey, Chair
ALA
Annual, San Francisco – June 16, 2001 – 2:00-4:00 PM
Metadata and Subject
Analysis: On the Verge?
Introduction and
overview of the work of the SAC Subcommittee on Metadata and Subject Analysis
Diane Dates-Casey, Governor’s State
University
Reports of the Committee:
Subject
Data in the Metadata Record: Recommendations and Rationale:
http://www.ala.org/alcts/organization/ccs/sac/MetaRept2.html
SAC Subcommittee on
Metadata and Classification, Final report, 1999:
http://www.ala.org/alcts/organization/ccs/sac/metaclassfinal.pdf
Comparison of Subject
Treatment of Several Metadata Standards.
A comparison of how
various metadata schemes implement subject analysis.
Aimee Glassel, University of Wisconsin, Madison
Report: Comparison of Subject Treatment of Several
Metadata Standards
http://library.uscd.edu/~becky/metadata.htm
(Note: this paper will
be moved to the ALCTS website in the near future.)
Role of Technology in
Metadata Production : an example
Development of
Simplified Subject Headings for Metadata
Rebecca Dean, OCLC,
Manager, Metadata Analysis and Implementation Section
Implementations by
librarians
LCSH
and Metadata Subject Analysis at Brigham Young University
Kayla Willey, Brigham
Young University
Applying Subject Analysis to a Geographic Image
Collection
Frank Cervone,
Associate Director for Library Information Technology Services, DePaul University
Linda Morrissett,
Associate Director for Collections and Access Services, DePaul University
7:00-9:00 Hilton & Towers –
San Francisco Room
OCLC breakfast (Cataloging)
See http://www.oclc.org/strategy/cataloging/
See meeting notes
Medical call number has
been added to CatExpress
New version of CatMe and
spellchecker available
Collection sets include NetLibrary,
Jstor, etc. http://www.stats.oclc.org/wcs_list.html
MARS (Marc record services) includes
authority control
Automated Collection Assessment and
Analysis Service could be used to convince faculty, legislators, etc. that we
have pressing needs
FirstSearch will have per-article
purchases in Sept.
SiteSearch support through 2002 then it
will die – trying to move to totally web-based services, i.e. WebExpress – I
went to the booth later and talked with an OCLC rep about what they plan on
doing with the digitization and archival storage that had been part of
SiteSearch – he has my card
Illiad is OCLC’s new ILL web interface –
Cataloging and metadata will include
previews and links to contents
Passport dies at the end of 2002; CORC
and CatExpress will be on a metadata desktop interface
Digital Collection Management and
Preservation - will build a digital repository; they are cooperating with Aleph
(ExLibris !! ) software to digitize historical newspapers
CDRS – collaborative reference with LC:
http://www.loc.gov/rr/digiref/memb-enroll.html
I am concerned about how they will
be merged CatExpress and CORC
Talked
with a woman from Wayne State about how to get reference librarians to submit
sites for cataloging
9:30-11:00 Grand Hyatt -- San Francisco Room
CORC
Users Group -- General meeting
Went
to SAC instead since I will be on SAC next meeting
I talked with Suzanne (president) and told her to keep up website for those of us who can’t get to meetings very often
http://www.sil.si.edu/staff/CORC-User-Group/DraftCORCmission.htm
9:30-12:30 Canterbury -- Garden/Lanai
SAC --
General meeting
MARBI report from last ALA re: SAC
#x and #2 for taxonomic
hierarchy approved
2001-05 is 4 proposals: “i” for integrating resources leader 07,
008/18 frequency includes “k” for continuously updated (853-55 also), type of
serial will be type of continuing resource with new types l=loose-leaf,
d=database, w=website
508 will be repeatable
for the use specified in the proposal
see
MARBI page for minutes and decisions
Sears Subject Headings
New
update of Canadian companion is just out
New ed will be 2003, ed. 18; in the
mean time 17th additions available electronically
Library of Congress liaison – Lynn
El-Hoshy
Library of Congress will sponsor a book
festival in October with the First Lady Laura Bush
Budget for 2002 is under consideration.
There is a net decrease of $68.4 million or 13.4%. The decrease is because last
year had a one-time increase to do a digital repository and preservation. Still
7.2% less than in 1992. They have achieved efficiencies and can’t get any
tighter so they are requesting money for new staff, especially for online
resources
Voyager - Waiting for a stable version,
hopefully Feb. 2002 to install 2000 version of the software. They still can’t
do authority records via the web and Z39.50 – just waiting. Still working on
serials checkin. They started trying to do titles alphabetically, but are now
focusing on special needs. They are trying to figure out how to do inventory.
The are manually working on fixing Afro-American subject headings because
Voyager can’t do anything automatically.
See her report for new CIP features. http://cip.loc.gov.cip/
CDS has allowed posting of PCC
documentation on the PCC website. The SACO manual with lots of good examples is
available (I have already cataloged)
New MARC21 data elements in LC bib and
authority records that will be present in records distributed by LC will be
listed at: http://lcweb.loc.gov/catdir/cpso
They will be setting up “Developed
countries” as a heading
The LC outline on the web has links to
greater detail. They use GroupWise and Wordperfect – if you don’t have it, you
can’t read the documentation.
Ethnic groups workshop was held.
American ethnic groups are direct order; others are inverted. A guest asked if
there couldn’t be more clarification on this issue.
Library
of Congress Action Plan
http://lcweb.loc.gov/catdir/bibcontrol/draftplan.html
SAC is mentioned as a possible
collaborator several times (2.2 & 3.2) and seemingly missed on 6.3 – FAST
“simplified LCSH” or additional things like getting rid of exceptions and
inconsistencies
SAC – semantic interoperability will
probably be a charge of some sort tomorrow
3.2 is exactly what Lois and IFLA are
working on
2.5 LCSH free on the Internet – fight now
faced with cost recovering requirements. Mary Charles Lasater said there was
discussion at Big Heads about doing subject analysis any more, yet the metadata
program showed how much of the work on the web is focused on some application
of subject analysis
Ann Taylor is doing research on
“keyword” searching – which really searches 6XX subjects and hopes to show
zero-hits and poor results when you can’t search 6XX’s. Wilson will be doing
relevancy ranking based on 6XX’s
David Miller said his stats show the number
of searches using subjects went up nearly 90% and zero-hits dropped to almost
nothing
(left for program)
11:00-12:30 Marriott -- Golden
Gate Section C.
"The Serials Pig in the Aggregator's Poke, the Sequel" –
ALCTS
Serials Section Policy, Research, and Publications Committee
(missed a little in
travel)
It requires LOTS of staff to get
bibs into the OPAC for aggregated titles.
Why in the OPAC? – need them there so
users KNOW we have the titles and do appropriate resources sharing (and also
adds local holdings links)
We need help – e.g. getting bib records automatically when you subscribe to an aggregator’s records, or more from OCLC Collection Sets, etc. Some are using SerialsSolutions to get bib records in OPAC; we need decent displays from vendors
Title-by-title cataloging needs to be done, but each library shouldn’t have to do it. Yet we can’t trust the vendors to give us good records because they haven’t always
We need to find a way to share work
Then there are things like Project Muse where there really isn’t an aggregator; they are a package. They are more reliable than aggregators whose titles come and go. Maybe that is where we should start
Single vs. separate – you can buy separate records, you end up with lists in your OPAC but there is a lot less time on maintenance; or, with single each and every library spends time editing the records every time there is a change
University of Lexington, just decided it won’t do aggregator records – no specific reason
One speakers shays she keeps track of fund (dept) records associated with journals to do reports, etc., so she favors separate records
Integration of records
Separate records out to be able to be brought into the ILS and the ILS should be able to logically display clearly for user
Get your vendor to use the GMD
One speaker doesn’t think young users who are used to web searching are willing to keep trying – they are used to clicking a lot on the web til they find something – we underestimate patrons
One library is using single records, then cheating with obsolete fields in holdings records to show which aggregator they use – problem is you can’t then do resource sharing (it might also affect things like Z39.50 searching)
“Embargo” titles – when articles are installed and active can be 3 days or 3 months. We have to demand to know if that is what is happening on a title-by-title basis. Unfortunately it is in the best interest of the subscription vendor to actually negotiate embargos
One library spends three hours a day
just checking Lexis-Nexis titles to see if they work
SerialsSolutions person
http://www.serialssolutions.com/Home.asp
They have 166,000 holdings records
available (won’t help with PALS unless there is also a brief bib) for aggregated
titles.
There
are errors and changes. How do we keep up???
One library said 75% of their ILL’s were for titles they had. Now that they are in the OPAC, practically none are requested
LibLicense is working on a way to put information on the web about ILL restrictions or conditions – of course each library may have negotiated something else. This seems to intense on a per-record basis – maybe on a record for the aggregator
A few libraries are clean-sweeping
their bibs every 2 months using SerialsSolutions, then reloading updated files.
Minimum level records are better than nothing
TDnet.com
is similar to SerialsSolutions
Why can’t we have lists that are current from vendors? Both for users and technical services. Users don’t like search boxes as much as they like a browsable list
Several said they don’t like it when records that lead only to table of contents or abstracts
WebFeet does cross platform
searching between OPAC’s and Web databases
2:00-5:30 Marriott –
Golden Gate Sect B
Impact of metadata on searching the Internet - ALCTS
Networked Resources (Clifford Lynch, Caroline Arms, Sally McCallum)
Clifford Lynch
Metadata isn’t useful by itself. It
has to be tied to a function
We have spent a lot of time figuring
out how to create metadata; but not the “how it works” – e.g. union catalogs
between multiple libraries, FTP’ing files, etc.
We need to learn how to move/manage metadata. Open Archives Initiative is an example; Dublin Core is its basis. It places little constraint on what is in the records
Web search engines
Deal with static documents
Deep web/dark web – the stuff in lower levels, like in library catalogs or dynamically created web pages – all this is lost
Open Archives – provides a standard way to get at metadata for stuff in the deep web or in protected databases.
ľ billion web pages exist but there are only 50 million unique titles
Do we really want to catalog each document, or can’t we catalog the database? If we rely on content-based retrieval, how do we get it to go into the deep web?
Metadata
It is a third party statement about an object.
We have pride in good cataloging – comprehensively, accurately created; whereas there are metadata out there that is purely advertising, spamming etc. How do we edit these out?
Open Archives has the capability of including evaluative comments that can be used to “filter” out poor sites that respond to a query – we have to look at provenance, authenticity, etc.
Caroline Arms – Metadata Harvesting to
Expose Treasures from the Deep Web
Why
we need metadata
Our
role is to bring together like items and differentiate among similar items
FRBR - Functional Requirements for
Bibliographic Records http://www.ifla.org/VII/s13/frbr/frbr.pdf
Svenonius, Elaine. The intellectual foundation of information
organization OC 42040872
– finding known item, identifying,
collection, act of choice, obtain, access, support navigation
Economics issues of metadata
Costs and motivation of
information providers
Technical efficiency
Description is expensive
to create and maintain
Hard to afford doing
more than once
Different practices are justified by
different classes of content and demands of primary users
Consistency of metadata
(lean vs. rich records)
Search tools matter
For
structured data SQL, relational databases work well
For regular
websites you have no structure
Identify
what works, fix what’s broken
What needs the American Memory
project found:
Users want to use to
search: name, geographic, date of coverage
Finding and collocating
by topic is very problematic:
Controlled
vocabulary vs. familiar terms (i.e. use both)
Need to
retrieve through general AND specific terms
Domain name,
specific vocabularies were important
Need things
like “more like this”
Biggest complaints
Too much stuff, yet
missing quality resources
How do we exclude or
provide filters
Need broad categories
Need to find
categorizations that work across tools and access channels
Learn from the marketplace in metadata
practices – is more consistency work the cost? Need Authority record control
Open Archives
Easy to join
Low barrier to entry
All kinds of media
Form of metadata needs
to be better
Benefits of metadata
Some records are already
harvested
We’ve tested cross-walks
– they work
We’ve done test mapping
– they work
Sally McCallum
(Library of Congress) – Exploration in
Control and Access for Networked Resources
Web objects and the need to save
Libraries do not collect
or control everything
There is the “ephemeral
web” and the “research web”
Ephemeral
Tend to be current
access materials; will they be archived?
Tend to be full-text uncontrolled
keyword searched but these usually isn’t any metadata to help searching
Ranking is undisciplined
We need metadata with source
indicators, a simple set of elements, minimum content rules, transformable
tagging, and encouragement for authors to use – Dublin Core is close to this
We
need to do something about the research web
Libraries select, pay for, etc
Libraries pay attention to, look to the
future, look for persistence, preservation, etc.
Can MARC content be simplified?
It
is cataloging-rule dependent (e.g. AACR2)
It
includes intentional duplication
Has
extensive parsing
Has
many data elements
Web documents have characteristics that
make a difference on the cataloging process
The
transcribed “title-page” is too slippery
Version control and periodicity are so
complex – do you need more or less data elements?
New techniques for retrieval might say
something about what cataloging you don’t need to do anymore
Open Archives Initiative supports XML,
full-text tagging, roundtrip markup
Cataloging is not monolithic now;
all information doesn’t have to be treated alike
There are important reasons to look
at new ideas
We need to apply collection policies
to electronic documents
The research web documents are
candidates for MARC but:
We need to review MARC
content or cataloging rules
Need to make structural
tools available
Chris Sherman – Dark
Web
Metadata is a wonderful concept –
will it ever be used?
Commonly there are 2 basic meta tags
<keywords> and <description>
RDF holds promise for realizing the
semantic web
“black hat search optimization (bad guys) actually utilize metadata or insert metadata to make web sites rank higher
Metadata’s problems and why search
engines don’t use it
Hotbot does use it
Alta-vista is using only
description
Google will never index non-visible
stuff because metadata can be set to useless data by the bad guys
There
is “cloaking” – bait and switch “IP delivery”
Web
servers can return anything from any page to a crawlers request
How
can we make crawlers use metadata?
Promise
implicit data – i.e. things like domain name reliability
Citations
– in a research arena, who listed the site
File
type can be specified/identified
Dynamic
database shields have partial names
Voluntary
standards won’t and haven’t worked
Metadata creation and validation
services that search engines could trust and “fingerprint”
Collaborative searching – metadata
works well in a closed system; so if, crawlers could cooperate with it, it
could identify it as “valid” and do the search
Pay-to-play – e.g. Inktomi has
significant improvements in index quality but it costs to come to the good
site. Certain sites are “tax exempt”, e.g. edu gov
Can set up peer-to-peer content
repositories and they agree to certain rules
Prognosis
Metadata improves
searching
It will fail until we
have a semantic web where we can communicate
4:30-6:00P Moscone Convention Center – Room # 224/226
Ejournals and the Web: Standards for Tomorrow – NISO/BASIC program
Presentations: http://www.niso.org/Ejournal-web.html
Time to balance the
needs for standards and the needs for innovation
Brian Green –
EDItEUR
http://www.editeur.org/onix.html
EDItEUR is the international umbrella body for trade standards organizations, libraries, book sellers, agents, publishers, ILS vendors
ONIX – came from the trade side
MARC is unfamiliar and
not useful to publishers
No equivalent standard
to mimic
Publishers not best know
for providing accurate bibliographic information
Standard for communicating
rich metadata
An XML DTC
International
collaboration of EDItEUR, AAP, BISG, BIC
Contains structures for bibliographic
information, reviews, images, links, audio, and video clips, price,
availability, rights management
For serials – journal
issue article, detailed bib info, A&I services, etc.
US & UK getting
close to using it
ISBN has never had a
metadata set – hope to have a core ONIX set it could use ONIX library sector uses
Selection, awareness,
services, enriched OPACS
Interest by both
national libraries to use and map – LC and British Library
National libraries are
participating in standards committees
Libraries should embrace
ONIX as a source of bib data
Library systems should
exploit in OPACs, Acq, etc.
Metadata without content
rules is not useful in the library environment
Publishers should work
with libraries to minimize discrepancies
Current developments
Moving from the book
centered, to variety of formats
Extensions have been
added for national needs
Video/DVD draft to be
piloted
Subsidiary rights
ONIX for serials
Surveyed lots of agents
– got 35 detailed responses
Identified potential
business applications
Enriched
cataloging information from publishers: images, price/territory
Structured
multi-level bibliographic information in standard XML format
Shipping, cehckin and awareness info –
let libraries know hen a release is available
Library
holdings with standard formats that link to appropriate copy
Building
block for usage statistics
Working to finish XML
and DTD, do pilots – TALK TO YOUR VENDORS !
Eric Van de Velde
What is Open URL? Transportable metadata format (embedded in HTTP GET or POST request) between information services (e.g. from web to document delivery)
It is composed of a resolver and an
identifier – http://server.name ?id=identifiers
Why do we need it?
Persistent
Embraces DOI which
provides persistence
The server address can
be changed and the identifier remains the same
If you aren’t authenticated for
full-text, you can still use the open URL to do document delivery
You have context sensitive links – the
link knows something about the identy of the user and can determine appropriate
copy
Open URL can log transactions of a user
if using same resolver (i.e. your configured server) and then allows you to
create something from your transactions, e.g. create a bibliography
A way to separate junk web from good
web, yet you don’t have to use static, passive URLs we are currently using
OPENURL listserv http://library.caltech.edu/OPENURL
See also: http://www.dlib.org/dlib/march01/vandesompel/03vandesompel.html
Herbert Van de Sompel herbertv@cs.cornell.edu & Oren
Beit-Arie oren@exlibris-usa.com
Debbie Loeding –
Wilson – Aggregators
Wilson indexes ca. 3000 titles, 1/3 publishers grant full-text, 1/3 don’t, 1/3 don’t even reply
Publishers sometimes don’t own rights
to an article in a journal, hence the aggregator can’t reply
Academic publishers are beginning to
demand 2-5 years embargo before articles are put up
When publishers denies full text, they
fear: print cancellations, low royalties, lack of control; they want statistics
on number of articles read; or they have their own service
Publishers often terminate access
because: they didn’t get enough royalties, developed their own service, fear
losing print subscription, or the journal changed ownership
Wilson does have arrangements with:
Academic, Ingenta/Catchword, direct to publishers, etc.
Audience questions
Could ONIX be designed to
standardize information about cessation, remote access, etc.?
How do we get ONIX-rich records for our catalogs? – on the BISG page there is a list of publishers – still not very stable, but maybe soon
These ONIX records could be like order
records that either themselves will be updated or replaced
ONIX is such that content could be very
different but structure will be the same
7:00-9:00 Hilton & Towers –
Continental Ballroom 1-3
PCC - Program for Cooperative Cataloging
Training manuals on
various’ sections websites will be put up. http://www.loc.gov/catdir/pcc/automation.html
OCLC Batchloading
Task Group is trying to find ways to make it easier to upload to OCLC
Task Group on Educational Needs of Catalogers – since so many library school have ceased teaching cataloging this has become a priority
British Library will begin using MARC21
when they bring up their new online system
Coordinators: Ana Cristan – BIBCO,
Carolyn Sturtevant – NACO, Ruta Penkunias – SACO
LC’s draft action plan is out for
comment
Impact of Core Records report on users
is due in Aug. They will be also be reviewing various core standards and
comparing them to each other in a chart
Automated classification demo will be
held at Midwinter for vendors
AACR2 Chapter 12 on Continuing
Resources is nearly ready – will be out in 2002; in the mean time they are
working on a new CONSER manual
Holdings – vendors seem to be trying to
apply the standards or adding 891’s. Diane Hillman will be working on “what does
it mean to be compliant with MARC21 Holdings Format”. There are 41 CONSER
libraries
Utilities Wish List panel http://www.loc.gov/catdir/pcc/pccpart01a.html
NACO – make it easy to create or authenticate bibs with NACO records in same system
SACO – streamline
Records
in natural language or translate into English – why should we have to
translate?
NACO – Z39.50 – if the ILS can support
OCLC ID #, OCLC developed a facility to create a skeleton NACO record, but most
ILS’s can’t do it
Multi-view records – could choose to see bib or holdings or bib-and-holdings record for download – Pattern Initiative is looking at developing a sharable holdings record
“bib notification” of changes to CONSER records – is that possible? Could 891 be a possibility?
OCLC and RLG are looking at FRBR principles for displaying works, expression, manifestation, item; PICA and some European ILS vendors have already looked at it – (FINALLY! – how long have I been saying this?!) Unfortunately there is little in most bibs to assist in the process – (but I do it whenever allowable with linking fields). Most bibs are “expression” or “manifestation”, the authority record may be the work, and the local system has the “item”
See section reports of LC’s ALA briefing for things discussed at ALA (CDRS, ILS, Digital Library, Cataloging Directorate): http://www.loc.gov/ala-update-2001.html
MONDAY – June 18
8:00-9:15 Hilton & Towers –
Continental Ballroom 4-6
CIS Breakfast
The
usual “ad” for CIS-LexisNexis and awards were given. The speaker Mara Liasson, White House
correspondent for National Public Radio was very interesting.
9:30-11:00 Moscone Convention Center -- Room 228
Copy Cataloging Discussion Group --
LC update, PCC core level records, CORC for copy catalogers – Lois
Shcultz, Head, Monographic Cataloging Section, Duke University and Allene Haye,
Senior Cataloger of Electronic Resources, Library of Congress, and LC CORC
coordinator
Core record
discussion
Core was being used by some libraries present. Most
said they really ended up doing full. Some complained that core records don’t
have enough subject headings.
You can use the core standard, but if you are not a
BIBCO library, you can’t fully code the record so other libraries know. 042 + 4
= BIBCO with authority records, 4 without 042 = core but not necessarily
authority records, 042 + I = full and authority records
Subject headings – core doesn’t mean you can’t have
multiple subjects – what they say is they will spend less on description;
however, they are not putting subjects on fiction. We should upgrade records if
there are no subjects
BIBCO has not grown much – there is a perception
that it is elitist; 1/3 of BIBCO records are at Core level (4); the rest are
really full
Survey – managers are happier with the idea of core,
than catalogers who work with the records are. Managers see it as a way to do
speedier cataloging, but catalogers see things missing that take time – if
those missing things affect searching, patrons lose out too
It seems there is need for more and more subject
searching, why should we want to decrease subject analysis?
CORC
Costs are the same as regular OCLC
beginning 2001 summer
Will replace passport
90% of Internet resources selected
at LC are already on CORC
Easy export feature and works well
with Z39.50 – depends on ILS you have
LC is using CORC as a reference tool, especially
authority searching for reference staff and copy cataloging since Voyager
doesn’t work well
Hard to get catalogers to look at Dublin Core – it
started out simple and now is very complex
Using it to update their URL’s – as a notification
feature
Reference staff select the free resources, paste URL
in the box, and click create; the harvester starts the record; the reference
librarian writes a summary note and a few keywords for subject headings; they
actually submit the record to OCLC (the rest of us probably wouldn’t do that);
the reference librarian emails the cataloger who is a traffic cop who funnels
the records to appropriate personnel. They use email
LC is editing in Voyager rather than on OCLC and
they aren’t always doing subject analysis. Because OCLC picks up Voyager’s
records daily, the upgraded record gets put back in OCLC and replaces the bad
one (they say L&N until then)
If you select resources carefully, sites rarely go
away although URL’s do change
LC has a project to archive so they won’t lose
access
LC Policy Office has said if you go into a record,
MAINTAIN it
To say LCCopyCat, LC has to start with full level
records, not even “4”
How LC uses CORC is driven by their internal
structure, their ILS and OCLC
LC is using OCL’s course on Internet Cataloging to
train people
One library said their reference librarians, start
like LC with the URL and summary note, then add 590 with a note to say catalog
or catalog and put on local webpage, then save the record. Some reference
librarians are learning more about cataloging because they like to see what
their record ended up looking like.
BYU, even though it is a BIBCO library, uses
full-level cataloging except for foreign literature
Fla. – they
see very little time savings in using core, and end up doing full most of the
time
LC is doing Core for electronic resources
12:00-1:30 San Francisco Marriott –
Yerba Buena
OCLC President’s Luncheon
Sat
with Gary Johnson. We discussed ILS vendors, RFP, etc. Luncheon and
presentation was nice; outlined OCLC’s Global Strategy.
1:30-3:30 Argent -- Boardroom
ExLibris
(Some notes)
Oracle – relational database –
windows – web
Strength – consortia-based
GUI, user web, Z39.50,
X-service each have their own servers to connect to API, then to I/O, then to
RDBMS with ODBC connections. Server for each interface feeds to same
application software
Bibliographic records
(YEAH !!) = A bib record is stored for the consortia. Each library has an
administrative record (ADM) That can have local notes, subjects, call numbers,
856 and in my library I see my stuff. MnSCU/PALS is going to have an
equivalency table that will allow only certain libraries to change the main bib
– still each library can have its own display – YES YES
Uses style sheets for local
definition of OPAC display
Relevancy ranking can be defined by which kind of
tag you want to use, e.g. title, subject, etc.
If no author in a browse, there is still a response
and the authority record shows up; it is clickable to the authority AND/OR does
a new search on the right name
Has pre- and post- filter
Electronic links – to remote or
local servers where I have scanned in an image
Series search works like it should –
LOVE IT!
Login – personal display – might
limit this to staff and faculty
Login necessary to place holdgs
Metalib & SFX
Metalib is a portal to
bibliographic records plus everything else you have acces to – other OPACS,
journals, etc. If you come in on a guest login, you won’t get to licensed
databases.
Metalib could replace
SiteSearch functionality
Endeavor goes to Elsevier
– not everywhere and anywhere you define; Endeavor can’t sell separately
XML, MARC, Dublin Core records work
in the OPAC – YES !
Public and school libraries are not who ExLibris has
marketed to, but they are working on a kids interface
Circulation – uses Windows environment; toolbar has
nice icons for ease of use; it is the Circ view of the database, not a Circ
module – YES ! (because this affects password control and workflow issues); you
can have a loan and return screen open at the same time
When you buy Aleph you get the whole integrated
software
You can key in OR scan the barcode and it has
automatic check-digit calculation – YES !
Patron file is systemwide, but can set up privileges
differently at different libraries and even within a library – all on one card!
Can have multiple addresses and set that information
Can choose to display checkout history or not; can
choose to keep only last patron
Homebound not fully implemented; pieces exist in
circ
Can print receipts, change due date, pay fines, add
notes, etc.
Have a basic media booking; it is being enhanced;
can request use of media, room, equip within a 2-week calendar.
Can set up a class of patrons that can place holds
on items on the shelves; default is that you can’t place a hold on a shelved
book
Staff can re-arrange order of holds
Canned reports – many
Loaders available for MARC, Yankee, Marcive
Still working on a customizable default workscreen
(LAD) for Acq
Working EDI with various publishers
Serial checkin has a pattern database and uses 891
to convert to 853-5
Implementing ISO toolkit for ILL and is in
partnership with MnLINK to do ILL
Upgrades are annual
DigiTool – stores data in Oracle
They have a link to chat reference – LivePerson
Can easily upload/download OCLC records
Demo staff well versed in product. Seemed to provide
“honest” answers as best as vendors do.
Ex Libris made a lot of great announcements at ALA
Product News - ALEPH 500
Ex Libris (USA), Inc. and ABLE(tm) Partnership to
Develop Bindery Information Interface.
http://www.exlibris-usa.com/news1.asp?categoryId=116&admin
Ex Libris Announces E-Reserves for ALEPH 500.
http://www.exlibris-usa.com/news1.asp?categoryId=121&admin
Ex Libris Taps TLC's Interlibrary Loan (ILL)
Software Toolkit to Deliver ILL Management Application
http://www.exlibris-usa.com/news1.asp?categoryId=118&admin
Product news - SFX
Infotrieve Teams with Ex Libris for Document
Delivery.
http://www.exlibris-usa.com/news1.asp?categoryId=115&admin
Ex Libris (USA), Inc. Announces Multiple SFX Sales.
http://www.exlibris-usa.com/news1.asp?categoryId=120&admin
Company News
University of California Davis Signs Up for ALEPH
500, MetaLib and SFX.
http://www.exlibris-usa.com/news1.asp?categoryId=119&admin
PTFS Signs Up as Federal Government Distributor for
Ex Libris (USA), Inc.
http://www.exlibris-usa.com/news1.asp?categoryId=117&admin
Ex Libris (USA), Inc. On Schedule With Increased
Staffing Plans in North America.
http://www.exlibris-usa.com/news1.asp?categoryId=114&admin
Ex Libris (USA), Inc. Announces Staff Promotions.
http://www.exlibris-usa.com/news1.asp?categoryId=113&admin
Ex Libris Ltd. Announces Major Appointments: Marc
Daubach to Corporate VP International Sales; Julie Booth to Managing Director,
Ex Libris UK.
http://www.exlibris-usa.com/news1.asp?categoryId=112&admin
3:30-4:45 PARC 55
EpixTech (Horizon)
(some notes)
You have to open up modules
one-by-one; have menu launchers
Three levels of security – network,
sybase (stays in memory), subsystem password
HP servers; Sun is their development
preference or IBM; Unix
Language: C++
Box can be brought up in windows
client to sort by things you choose
Can use Z39.50 to search other
catalogs or import records
Not aware of using multiple holdings records for
electronic journals – (they should be! See ISSN section of: http://www.loc.gov/catdir/pcc/bibco/opco01.html
) It just dawned on them that you have to be able to circulate NetLibrary files
and/or rocket readers
Have to do setup for each Z39.50 client; they train
us, then we fix – (is this how they would expect we approach our consortia??)
Cataloging templates – have brief help at bottom but
not even as much as MARC21 Concise – (poor !)
Global change is possible; you always have to start
your catalging windows over again to see what changed
Authority – it links if you want it to, but we
couldn’t see the record (bad! It also worked under the assumption that all
people in your catalog with a particular name were the same person – since we
haven’t done authority work on every single entry – nor would we - that will
never work!)
Item record – 5 screens, with 1st
screen showing barcode, copy, call number and volume; other information is on
following screens – (annoying clicking – staff won’t want to have to click that
much to see info)
Barcode is 14 digits but doesn’t have
check-digit – their autocopy barcoding could never work
Checkout can be over-ridden by another
person; if they change password, it would remain on the session; so you have to
have another login window to override – (when they showed it, you couldn’t tell
which of the 2 logins was displayed, so the wrong one could very easily be used
unknowingly)
Two notes are possible on an item, each
255 characters
Borrower types – in a consortia you’d
see every library’s – could have hundreds you’d have to pick from. They could
be organized logically, but if you are the last library on the list, you have a
long way to scroll down to find your type just to set up your patron
Patron records – have permanent and
temporary address with dates valid; the have a place for email address but the
demo staff didn’t know where the notice would go if the email were wrong –
(bad!)
They use Teleser for autophone notices
but they have problems when they encounter DSL lines – probably true for anyone
Homebound – keeps reading history
Ended short of scheduled time because
“next group was there”. I didn’t see them and we missed out on 15 minutes of
our scheduled time. Demo staff didn’t seem well versed.
TUESDAY – June 19
Visited with several
vendors
LC – to get a copy of my quote they
are using to advertise the classification web
OCLC – to discuss the migration off
SiteSearch for digitizing as well as impact on LaND
Gaylord Polaris – brief demo