|
FAQ:
Retrospective Conversion
Barbara Leff
Retrospective conversion, commonly known as "recon" or "retrocon,"
means to convert a manual card catalog of paper records into machine-readable
online computerized records.
The recon process can be done in various ways: One is a very lengthy
process—to enter the cataloging data for each library item directly
into your online cataloging software, resulting in your own online records.
Others are faster and costlier, using vendors and/or your personnel
to match existing computerized records and transfer the results into
your own library online catalog. There are many steps in this process—this
is but an introduction to the topic.
The recon process has been described by various AJL members from one
extreme to the other—some say it is "fun,""enjoyable,""no big deal"—others
say it is "complex,""difficult," "I'm overwhelmed and don't know where
to begin!" Knowledge and attitude play a large role in how one reacts
to the recon process. There are many options in the recon decision-making
process, and many "right" decisions. Only by doing your homework and
more research, including looking at the references at the end of this
article, can you hope to choose the best, considering the time, money,
and staff available. If money is an issue, don't compromise library
standards—wait until you h ave enough funds to do it right.
Many AJL members have kindly shared their methods and comments—I give
them credit at the end of this article.
I was unable to find any one recon outline, from start to finish, that
covered all that I considered important, so I adapted my own, basing
some of it on the OCLC booklet referred to at the end of this article.
Here is an overview of recon steps—to give you a sense of the entire
process. These steps are in an approximate order—many are interconnected
with each other and are developed simultaneously:
- Decide whether or not to convert;
- Understand conversion process;
- Set objectives;
- Develop a plan; select mentors and/or consultant; anticipate problem-solving;
consider personnel needed;
- Secure full funding and/or firm pledges of funding for a phasing-in
process;
- Set up a systematic method of dealing with the process, documenting
it, and being consistent;
- Identify expectations and visualize end results;
- Determine requirements, priorities, and how to handle new acquisitions
simultaneously;
- Conform to library standards;
- Prepare database—weeding, inventory, matching records, cleanup,
local fields;
- Select automation software, if one not already in place;
- Order barcode labels and reader;
- Prepare online catalog for Web (Internet) access;
- Select independent or group recon (and automation) options;
- Collect library data;
- Consider non-roman languages (e.g. Hebrew, Yiddish);
- Choose recon method;
- Set up system for error det ection, conflicting records, and authority
control;
- Choose whether to close or freeze card catalog while recon in process;
- Set realistic time schedule;
- Schedule staff and patron training.
Here is the recon process, as I see it:
Decide whether or not to convert
Very small libraries with minimal or no circulation, little growth,
and little use, can manage with a manual card catalog quite well. The
rest should seriously consider recon, for the same reaso ns that one
decides to automate a library—all discussed in detail in my previous
article on Library Automation—"SSC Gets Wired," in AJL Newsletter,
November/December 1997, p. 20.
Understand conversion process
Review the entire recon process so that you truly understand it,
especially the "Choose Recon Method" section. Research the resources
listed at the end of this article and ask vendors for recon literature.
The more you understand, the easier it will be to do all of the rest
of the steps in this process.
Set objectives
The main objective—the quality and integrity of your online catalog—should
guide you as you set your other objectives. If funding is very limited,
and time is short, then postpone recon—otherwise you might find yourself
cutting too many corners, resulting in such a poor quality online catalog
that the entire recon project may have to be redone in the near future—wasting
money, personnel, and time. This has happened to well-meaning but short-sighted
librarians and administrators.
As you read these steps, you will see that there is no one way to do
the recon process. Recon is complex and is not a stand-alone process.
How you do recon depends on where you are with classifying and cataloging,
software and hardware, staffing and funding, weeding and inventory,
among other things. But don't be concerned. Think of where you would
like to be, set your objectives, then move forward slowly through the
steps, and it will all fall into place.
Develop a plan; select mentors and/or consultant; anticipate problem-solving;
consider personnel needed
After you read and explore ALL of the steps in this process, do NOT
move forward without detailed planning. Planning is the most important
part of this process—to save time, money, and frustration. Expect to
problem-solve throughout the project. Mistakes happen but don't agonize
over them if they do.
Try to find someone experienced, whose judgment you trust, to use as
a sounding board and for advice, e.g. librarians at local schools/colleges,
colleagues at AJL national conventions and via Hasafran (AJL's electronic
newsletter). Salespeople don't usually tell you all you need to know
unless you ask the "right" questions about a product or process. If
you cannot find library people with recon experience to serve as mentors,
then a recon consultant may be the best choice for you if you can afford
it. See "Choose Recon Method—Consultants."
As you plan, consider use of personnel. Even if you outsource the recon
project, you will still need onsite personnel to complete the job. If
you do recon inhouse, it lends itself to doing tasks in an assembly
line fashion. As for any library task, use your talent selectively—professional,
para-professional, clerical, students, and various levels of volunteers.
Those with computer skills can enter data and search the databases;
librarians can doublecheck the final records and correct the subject
and descriptive cataloging to conform to library standards; others can
do the varying levels of clerical tasks. Amy Wissoker Graham writes:
"I think retro-con would be a great project for a confirmation class."
Most importantly, discuss your plan with your supervisor so there will
be no surprises and you will have full administrative support.
Secure full funding and/or firm pledges of funding for a phasing-in
process
The more you understand about the recon process, the easier it will
be to calculate, compare, and explain costs when you need to find funding.
After estimating costs, add at least 10-20% because by the time you
get the money, prices will rise, your collection will be larger, staff
salaries will be higher, etc.
Funding usually comes from these sources: Budget request, grant proposal,
and/or donor solicitations. Try to enlist the help of your supervisor
or the fundraising arm of your institution for this purpose. Administrators
may apply for funding from sources like the Bureau of Jewish Education,
and library technology grants are sometimes offered by the state and
federal governments.
An overview of costs, which will run thousands of dollars, is too difficult
to determine for the purpose of this article because there are so many
variations and options. My best advice is to read through this article
carefully, get several estimates comparing the same services and prices;
then speak with your mentor or consultant to determine the best method
for your money and your library. Be certain to include funding for staff
and faculty training.
If you are phasing in recon, after a while it will appear that recon
costs and annual automation maintenance/upgrading costs for the online
system will merge, but they do not. Be sure to continue setting aside
funds for recon, to be kept separate and apart from your annual library
automation budget.
This recon project assumes that library automation is already installed
in your library. If it is not, however, combining the automation and
recon projects will work well and may be more attractive to potential
don ors, even though more expensive, because you will also need hardware
and software—tangible items which donors understand. (See my 4-part
Library Automation series in AJL Newsletter, starting with the
November/December 1997 issue.)
Set up a systematic method of dealing with the process, documenting
it, and being consistent
Set down a plan of action. Be sure to confer with your administrators
on what you are doing, your time frame, your funding, etc. Maintain
a file with your plan and record all the steps, your progress, your
problems. This will come in handy as a reference from time to time.
Be consistent in following the plan. Review past steps and look to the
future. Give each step a lot of thought BEFORE you take action. For
an efficient work flow, you might prepare a procedure manual and work
flow charts for some of the tasks.
Identify expectations and visualize end results
Talk with your staff, principal, teachers, parents, and students
to discover what they will expect from a computerized catalog. They
will have had experiences in public libraries and schools, professional
libraries, and may be able to give you a fresh perspective. Walk through
the steps of circulation, acquisitions, reference, retrieval—to be certain
that the newly converted data will meet their needs (e.g. children's
reading level, curricular topics). Visualize not only how the recon
work flow will all come together into one satisfying experience but
also how the end results of a seamless online catalog system will affect
your library. Do you think patrons could learn to like an online system,
as they used to love to browse card catalogs?
Determine requirements, priorities, and how to handle new acquisitions
simultaneously
If funding is adequate, convert all of your collection at one time.
This is the easiest, most efficient, and best way to go. Susan Dubin:
"We started to computerize our catalog ten years ago, but at the
time my synagogue did not want to spend the $7000 estimated to hire
a company to do a retrospective. THIS WAS A MISTAKE! We are still in
the process ... 13,000 of our almost 20,000 volumes online ... we are
going shelf by shelf and cataloging any books that are not yet online.
Because it is ... in addition to regular library work, it is a very
slow and tedious process."
If funding is inadequate for the entire project, consider phasing in.
Divide the recon project into doable sections with a 2-3 year plan,
e.g. identify and prioritize high-use collections or subjects and do
th ese first. For example, if Bible, Jewish holidays, life cycle, and
Jewish history are the highest use, then convert these first. Another
priority might be circulation when doing an inhouse conversion. Susan
Dubin: "At first we cataloged the books that were checked out,
as they were returned. The advantage to this method is that we had the
most popular books online and accessible at an early date."
Another viewpoint—if you have items that are in the various stages
of cataloging—from fully ca taloged to brief records to uncataloged,
then the priority might be to convert the partially cataloged and uncataloged
first; then the fully cataloged items.
With regard to new books and materials, as you acquire and catalog
them while recon is taking place, be certain to conform to the identical
standards and guidelines as those to which you are converting your records—so
that all will be the same when the project is completed.
Conform to library standards
Each online record is like a catalog card, but online records provide
greater access than a catalog card. Library standards are like the glue
that hold your online collection together; and future goals may depend
upon these standards being in place. For example: Library standards
will allow the flexibility of a new automation software/system if you
outgrow your current one. Also, you will be open to cooperative ventures
with other libraries in these days of shrinking budgets. Standards imply
quality—and you'll be proud of your library catalog quality when everyone
views it on the Internet, as part of your institution's Web site.
Library standards are: MARC (MAchine-Readable Cataloging) format, AACR2r
(Anglo-American Cataloging Rules, 2nd ed. revised), classification schemes,
subject headings, name and subject authorities, and Z39.50 Web search
protocol—covered in "SSC Gets Wired" in AJL Newsletter, May/June
1998. Do not compromise on cataloging standards whether your recon is
inhouse or by a recon vendor. (See also "Prepare online ca talog for
Web (Internet) access".)
CASPR's LibraryCom now offers a new service called FreeMARC.
This allows you to download MARC records over the Web for free— from
1.8 million MARC records licensed from the Library of Congress, 1968-1999
records. [If you should use FreeMARC, do know that you will NOT find
MARC records for every title in an average Judaica library.]
Prepare database
• Weed the collection.
Whether your books and materials are cataloged or not, weeding your
collection is essential. Florette Brill: "...a good weeding should
be done ahead of time, since you will be paying for each book catalogued."
Each record of a book or material costs money to convert, store, and
keep current. Therefore, remove or withdraw all items that do not keep
your collection viable. Ask your principal and teachers for help in
weeding unfamiliar topics. Remove unwanted duplicates. Weeding (or deselectin
g) does take time, but it will definitely save recon money, and most
importantly, it will be a wonderful opportunity to get to know your
collection and put it in good shape.
• Inventory the collection.
If you have no inventory record (shelflist cards) because your collection
is uncataloged or poorly cataloged, then you can create an inventory
during the recon process—but only if you do recon onsite with items
in hand.
If your collection is partially or fully cataloged, then an accurate
inven tory of what you own and plan to convert must be represented on
cards or in your current online records. When doing an inventory, be
certain to include all items in your collection, including those in
circulation.
Lisa Handelman inventoried 10,000 volumes of a synagogue/day school
collection in about two weeks, with volunteers (Moms and teen-agers)
for full day commitments, a few assistant teachers, minimum wage teenagers
for 3 days to fill in the gaps, and more Moms for two hours at a time.
First when she weeded, she removed books but wasted no time making withdrawal
notations on the shelflist because it would be handled during inventory.
Then, to do the inventory, she made sure all shelflist cards were filed
and in proper order, and the same with the books on the shelves. She
divided the collection into categories, e.g., picture books, Elazar,
Dewey, Biography, Juvenile fiction, Juvenile Judaica fiction, etc.,
and placed the shelflist cards by category in labeled shoe boxes. One
box went to two p eople who were partners—one worked with the cards,
barcode labels, and some blank cards, and the other with the books and
label protectors. The book person called out the title of the book,
the card person matched the card and affixed a barcode label to the
card, and gave matching barcode label to book person to affix to book,
cover with label protector, and return book to shelf. If card missing,
card person filled in a blank catalog card with title, author, ISBN,
stuck barcode label on card, and filed card in proper order. If book
missing, removed card and place it upside down in a "missing books"
pile. If duplicate copy of book missing, crossed out copy number and
assumed book lost. If missing books/cards found, then put in proper
place. When done, cards representing still missing books were thrown
away. Lisa wrapped the shoe boxes filled with the cards, and sent them
via UPS, insured for $10,000, to the recon vendor to match records.
She supplied lots of coffee and doughnuts to her helpers to make the
proc ess enjoyable!
• Define the match and complete online record
All the information for matching what you need with what a source
database has, as well as for completing and upgrading your online record,
should be in one place—a single catalog card for each title you want
to convert, or a computerized record. This means that you should have
the following information on a single card/record for each title: Author(s),
title and subtitle, edition, publisher, date, unique number (i.e. ISBN
- International Standard Book No., LCCN - Library of Congress Card No.,
or ISSN - International Standard Serial Number), number of copies of
paperback/hard cover, call number, location (Ref., AV) and subject headings.
Most small libraries don't have all this information in one place. So,
you either take the time to make this available before recon, or you
take your chances on getting inaccurate or incomplete results, and spend
time doing "cleanup."
FYI—for searching a source database, the best matches are numbers,
i.e. ISBN, LCCN, or ISSN, because numbers can be quickly entered into
a search field. However, numbers can also be copied wrong—yielding the
wrong record. Titles alone are not a good match, because titles are
not unique. Title/author and date are good search terms but take a little
longer to enter.
• Establish cleanup method
"Cleanup" refers to correcting data in your records—the original
data, the source database, and the converted data. "Clean" data helps
meet your main recon objective—for quality and integrity of your online
catalog.
To illustrate the importance of cleanup—here's a true story: Two synagogue
libraries used the same recon company. One sent me a glowing report
of a high hit rate, the other described lots of problems and a low hit
rate. Their collections were similar. Upon investigation, I found that
one sent them "clean" records; the other did not. The lesson is obvious.
[A "hit" refers to finding a correct matching record (hit) for your
item.]
To create a "clean" record be fore the recon method—check content of
existing records, correct the data, check validity of data, and reclassify
as needed. Follow the guidelines in "Define the match..." above. Some
feel this is too time-consuming, that they would rather take their chances,
accept the problems, and correct whatever is necessary afterwards. If
you are outsourcing your recon project, many errors will be corrected
automatically when converting to MARC records; thus, cleanup afterwards
becomes a viable option.
Source datab ases also contain errors. Steve Bergson writes: "One of
the unhappy disappointments I have found is that there are errors in
some of these grand catalogs that should not be there." Corrupted data
in the source databases might include typos, misspelled names, extra
spaces between words, wrong birth/death dates, wrong call numbers. and
inaccurate subject headings.
To make cleanup easier after the recon method, you might do the following:
Ask recon vendor to separate the no-hit records which were keyed in
(often with incomplete data), and flag those records for your cleanup.
Have recon vendor supply you with smaller files of records—instead of
all 3000 converted records in one file, separate them into 500 or 1000
to a file—easier to handle for cleanup. Instead of importing all converted
records into your automation software immediately, import them into
an intermediary software, like MITINET/marc (see "Select automation
software," second paragraph), wherein your cataloger can clean up the
records. Then, export clean records into your library automation system.
If your no-hit records have missing subject headings, a volunteer can
copy them from your saved main entry cards.
If cleanup is not done before converted records are merged with new
acquisitions, original cataloging, etc., it will be nearly impossible
to find the errors. Steve Bergson suggests "Once a month, casually browse
the catalog's author, title and subject entries ... correct any errors
as necessary." During retrieval or circulatio n, you can check for errors
too. But it is best to do it "right" before merging, and be comfortable
that the integrity of your online catalog has been maintained. (See
also "Set up system for error detection...)
• Determine local fields.
Consider carefully the local fields, to customize records to your library,
school, or institution needs. Be certain your software and your source
database are compatible for this copy-specific information. For example,
if a synagogue/day school uses two classi fication systems—Dewey (for
secular) and Elazar (for Judaica), into which two MARC fields will these
call numbers go? and can your software accommodate this? and can you
make the necessary adjustments for these to go into the correct fields
from the source database? Similarly, local fields can accommodate reading
levels, curricular subjects, etc.
Select automation software, if one not already in place
Automation software was already covered in "SSC Gets Wired" in AJL
Newsletter, May/June 1998, pp. 13-16. For example, the best software
is an IOLS (Integrated Online Library System)—one that follows library
standards of MARC format, supports all media and resource sharing, allows
multi-tasking (use of computers with other software, CD-ROMs, Internet,
etc., without quitting IOLS software), and has the Z39.50 Web protocol.
Actually, you may make your decision on automation software, either
before, during, or after recon, depending on how quickly you want your
patrons to access your online catalog. Esther Nussbaum and Chaya Wiesman
state that they started the recon process last year but don't need the
automation software until their new school opens in the fall. Their
consultant recommended BookWhere and MITINET/marc software programs
as temporary substitutes for the automation software. [Author's note:
BookWhere is for Windows PC; a comparable program for Macintosh is eZcat.]
BookWhere is a Z39.50 search engine which allows one to search other
libraries' cataloging on the Web and to download their MARC records.
The saved BookWhere records are then opened in MITINET/marc where catalogers
enhance and edit the records, add local call numbers and data, delete
unnecessary data, etc., and save it for their automation software. This
software also automatically corrects punctuation. When they are ready
in the fall, they will select their automation software. They are hopeful
that one incorporating Hebrew records with non-roman data will be available.
Edythe Wolf had "many problems" when converting from two computer software
programs—from Librarian's Helper DOS database of 7000 titles, and from
Microsoft Access containing their patron file—to one IOLS, i.e. Winnebago.
"About 1,000 records didn't convert or needed a lot of work. We had
to hire someone else to convert our patrons to Winnebago. Winnebago
gave us a refund for the patron conversion program cost, that they originally
charged us." Conversions from one automation or database software to
another are often fraught with problems if one is a non-standard format
(non-MARC).
Some automation software companies handle recon. Susan Berson asked
good questions and decided not to use Winnebago for recon because they
had a much smaller recon database, and few older or out-of-print books,
than OCLC. These are two examples of a company with a good product (IOLS)
providing other services that do not satisfy clients.
Order barcode labels and reader
Barcodes identify a library item through a unique machine-readable
code. Then each code is linked to the online record. In this way, a
book and its record are linked together for cataloging, circulation,
acquisitions, searching, overdues, etc. Talk with your barcode vendor
about "smart" and "dumb" barcodes.
Barcodes labels/numbers are often affixed to the items and added to
the records as they are converted by a vendor or yourself. Sometimes
they are done in advance during inventory. Choose barcodes only after
selecting automation software to ensure compatibility with the software
; many automation software programs automatically assign barcode numbers,
helping you to avoid mistakes.
Some things to remember about barcodes: Barcode labels have limited
shelf life for the sticky side, so buy only what you will use in a reasonable
period of time. Barcode labels can have errors, e.g. the reader reads
a wrong code resulting in a erroneous transaction—so get the best barcode
quality you can afford. The easiest way to avoid problems is to purchase
both barcode labels and a barcode reader from the same vendor, then
that vendor will be responsible for all problems. Some libraries produce
their own bar codes with a barcode machine, with varying levels of success.
Prepare online catalog for Web (Internet) access
An incentive for upgrading your records and conforming to standards
during recon is that your library catalog may soon be out there in cyberspace
for the whole world to see—once it is published on your institution's
Web site.
Many institutions are now on the Internet and want their library catalog
available for teachers, students, and members to access from home computers—for
curriculum planning, research preparation, and the prestige it will
bring them. Also, publishing on the Web is a cost effective way for
an institution to bypass hard wiring of a campus network when the onsite
classrooms and offices have Internet access. When they do include the
library, be certain to talk with the technician about setting up a firewall
to protect your data. You don't want to take a ch ance that hackers
will corrupt your library records.
If your institution cannot afford the expensive hardware, software,
storage, etc., required for putting library catalogs on the Web, even
though you may already have the library automation software with MARC
and Z39.50 standards, there is another option—a new service. CASPR's
LibraryCom is a library hosting service—an inexpensive alternative for
putting a library catalog on the Web. You simply export MARC or other
text records from your current system and email or send them to LibraryCom's
Webmaster; your records are available on the Web within 24-48 hours.
Abigail Yasgur just called to say "It works!—her library catalog is
now on the Internet—published on the LibraryCom web site with little
effort, for less than $400 annually, and will be updated quarterly.
Her library's MARC records were uploaded onto the LibraryCom site. Access
is by name and location on the site, and there is a link from her
library's web page to her library
catalog at (Jewish Community Library of Los Angeles). She said that
it's an exciting and inexpensive solution for Web publishing, and makes
recon worth the effort.
Select independent or group recon (and automation) option
• Automate own library and recon on your own
Most synagogue libraries automate on their own, and follow through
with a recon project. Often, this meets the needs of their institution
and library better, even though it is more expensive and time-consuming.
But, perhaps it is time for our smaller synagogue, school and center
libraries to look for ways to share resources as well as our library
operations and services because of the high costs, limited budgets,
and emerging technology that can now accommodate more of our needs—as
described in the next two options.
• Join existing union catalog or start a new union catalog
Some communities already have union catalogs with online records
and shared resources, e.g. private or public school networks, Jewish
community library and synagogue library networks. With an established
network, there are less decisions and costs when your library converts
its records. Before joining, be certain it will meet not only the current
needs of your library and patrons but also your future needs.
CASPR's LibraryCom also provides an easy way to set up a union catalog
on their Web site.
• Establish a consortium and all recon/automate together
Consider joining together with other synagogue and/or day school
libraries, forming a consortium for many advantages, including planning
to do a joint recon (and automation) project. It will take a lot of
cooperation among libraries AND their institutions to develop a successful
consortium; large libraries with greater turf concerns are finding consortia
worthwhile. But again, a caution that shared costs and cataloging, among
other shared projects, are appealing as long as they meet the objectives
of your library and patrons.
Collect library data
Data and statistics are necessary to help make recon decisions.
Also, potential recon vendors and barcode vendors will ask for some
data. It will be desirable for you to have a somewhat accurate knowledge
or count about the following:
Kinds and amount of materials in your collection (books, audiovisual,
electronic resources); total number of titles in your collection; approximate
number of titles pre-1968; secular children's titles; Judaic adult titles;
Judaic children's titles; number of titles in English, Hebrew, Yiddish,
Ladino, or other languages; number of records with ISBN, LCCN, ISSN;
number of titles you want to convert in each phased-in stage of recon;
people and skills available to work on a recon project; number of potential
patrons; retrieval needs of your patrons; circulation statistics if
you're phasing in with circulation records first; classification scheme;
subject heading list; MARC format; automation software.
Consider non-roman languages (e.g. Hebrew, Yiddish)
The best source for non-roman bibliographic records (e.g. Hebrew,
Yiddish, Arabic), according to Rick Burke, is RLIN, the largest
repository and source for bibliographic records containing Hebrew vernacular
data (particularly authors, titles and imprints in Hebrew as they appear
in the original works). RLIN does not provide recon services, but RLIN
can be searched, at a greater cost than the average recon search fee
from vendors, such as OCLC. However, currently, only high-end expensive
automation software like Aleph, Innovative, and VTLS can accommodate
MARC records in these languages.
For smaller libraries, Rick does not know of a single simple solution
for managing roman AND non-roman data in the same database. There are
future possibilities, but nothing appears to be available that will
support the MARC format a t this time. Steve Bergson writes,
"I am among those who are getting more and more disappointed with Mandarin's
ongoing delay of the Hebrew integration into their cataloging software."
Esther Nussbaum writes: "None of the automation programs, suitable for
a collection of our size, can incorporate Hebrew language records. We
are therefore continuing with a card catalog for our Hebrew titles."
Rick mentioned the option of Hebrew transliterated into English to
serve the patrons of small er libraries. These records can be found
in the OCLC bibliographic database, and OCLC has special reduced search
rates for recon.
Choose recon method
The recon method you choose depends on funding, collection size
and languages, quality demands, time constraints, control of process,
cleanup timing, and staffing, among other things—and each method has
advantages and disadvantages. The decisions are difficult, but as you
learn about the various methods here and you research it further on
your own and talk with others, the choices will narrow until you will
feel comfortable selecting one or more methods. This is an overview
of the methods available: Full vendor contract (onsite or offsite);
vendor database for batched recon records; inhouse recon; consultants;
or a combination of these.
No matter what your choice, in order to conform to your local needs,
be prepared to do inhouse doublechecking and some cleanup of records
as well as original cataloging for those records not found in source
databases.
Details of recon methods are complex, and space precludes my attempting
to provide any more than an overview of the various methods. I am naming
here but a few of the many vendors and products on the market; inclusion
is as an example, not as a recommendation. You must do your homework
to check the track record, reliability, etc., of the vendor. Vendors
with a quality product one year may not have a quality product the next
year! Vendors named here provide many more services than those described
in the examples.
To learn of more vendors and products, see references at the end of
this article, and ask this general question to vendors: "Name other
vendors that do the same thing that you do and tell me why I should
choose your product or services over theirs." In this way, you will
be able to identify others, compare and contrast, and draw your own
conclusions as to the best for your library and patrons.
• Full vendor contract
Vendors can only estimate final costs based on approximate nu mber
of records, hits, time frame, etc. Adaire Klein strongly recommends
that the contract contain a penalty clause if the job is not completed
by the agreed upon date with a reasonable leeway (3-6 months).
•• Onsite and Offsite: Rick Burke is "working with Library Associates
to do a small recon project for the UJ's Jewish history collection,
because LA has a staff that is able to work with Hebrew (non-roman)
records; the approximately 3500 RLIN records will be uploaded to his
Innovative Interfaces online catalog."
Library Associates is an example of vendors who can do everything—from
start to finish. They do parts of collections or the complete project,
and fees are based on number and kinds of records to be converted, general
services, and special handling. With this method, your library controls
the project because of the specifications determined at the outset during
vendor's initial site visit, and the continuing interaction between
vendor's project manager and the library. The vendor maintains a full
technical services staff with professional catalogers, who can work
on your site or theirs; thus, you do not need additional personnel or
equipment. They will also provide training for your staff as needed.
Library Associates prepares "web pages for our clients to view as we
work on their various projects, so they can see and use the same tools
we use, and keep current on our progress." Looking to the future, Library
Associates wrote, "Increasingly we are being asked to catalog relatively
non-bibliographic materials, especially as libraries are asked to support
Intranet..." [Intranet is an institution's inhouse Internet.]
•• Onsite only: The vendor comes to your library to do the recon process.
As an example, Library Conversions brings a crew of 6 people (including
a cataloger), a network of laptop computers, a CD-ROM tower housing
almost 9 million full MARC records plus other bibliographic records
from many sources, special book trucks, etc., to handle 10,000-12,000
records in a week. Their major advantages are you don't need to do any
advance preparation, except weeding, because they work directly from
your books, materials, and catalog cards thereby reducing the margin
of error; they handle the work flow and use whatever call numbers and
subject headings are on your cards or they will take the information
from their databases; and you are there to supervise and answer questions.
When they leave, your online catalog is up and running and your books
are back on the shelves. Their fees are reasonable and competitive with
other recon methods even though also covering their personnel expenses
(travel, food, hotel). Their most reasonable rates are for Dewey Decimal
Classification System and Sears Subject Headings.
• Vendor database for batched recon records
After you "Prepare Database" (see above), you can send all your
shelflist cards to a vendor to convert. Marcive is typical of many such
vendors. Working directly from your shelflist cards, Marcive matches
your recor d with one in their database. Then, they can key in the local
data (call number if different from the database, location (Ref., AV),
copy/volume, barcode numbers, price, fund) for an extra fee, or you
can do yourself. The no-hits are pulled from the shelflist and bundled
separately. Marcive will key no-hits into the MARC format for you to
review, cleanup, complete, etc., or return them to you untouched. The
converted records—data processed in batches—are saved to a file, and
sent back to you on tape, disk, or over the Internet, along with your
cards. They can also print new spine/book labels. This is a common way
to do recon; and an easy way to phase-in by sending a thousand or so
cards at a time.
If your records are already online, then Marcive has a Brief Record
Upgrade service, converting your records to MARC, adding subject headings
and other cataloging data from their database, etc.
Susan Bergson uses the Winnebago library software and asked that company
to do the recon project. She wrote, " ...it was a relatively painless
process. Winnebago was great to work with, and they have been extremely
responsive to any questions or problems. They created records for the
no-hits. So far I have been very pleased."
• Inhouse recon
•• Key in records and do original cataloging
As stated before, you will need to use this method when source databases,
discussed in the next section, do not have a match for all of the items
in your collection, for whatever reason. This method is the least expensive
if your personnel are bright, capable, reliable volunteers, and a few
are computer-savvy with excellent typing and proofreading skills, and
all are willing to make a long-term commitment to the project. If you
do not have this quality or quantity of volunteers available, don't
attempt it. Also, don't attempt it if you have a large collection because
it will appear to take "forever" (often 10 or more years) and you will
be in a constant state of feeling disorganized, and your patrons will
feel it too, with no end in sight! (I speak from experience on this!)
To do this method, a data entry person keys in (types) data into your
automation software (adhering to the MARC format), by copying data directly
from the book or material in hand and from librarian's notes of classification
number, subject headings, etc. The librarian then checks each entry
for errors. The software program automatically assigns barcode numbers,
etc. This may be similar to your method for handling new acquisitions.
On the plus side, you have total control over the quality and content
of your online catalog. However, your catalog is a work-in-progress
for lots of years. A cataloger must be available to answer questions
and resolve the many problems throughout. Also, expect volunteers to
resign, requiring you to find new talented people and train them throughout
the project.
•• Access data in source databases
For this article, I've used the term, "source databases," to refer
to two categories: Bibliographic utilities/shared databases (e.g. OCLC)
and/or vendor databases (e.g. Marcive), which are fee-based; and library
catalog databases, which are viewed for free on the Internet (e.g. Library
of Congress).
Be sure to examine the source database itself before making your choice.
Know the kinds of materials in your collection and be certain that the
source database can provide it. The scarcity of Hebrew language records
has been discussed in "Non-Roman languages...". Most companies specializing
in school library conve rsions use Dewey Decimal and Sears Subject Headings,
and do NOT have a good hit rate for Judaica children's books and materials,
except those published by major publishing houses. Therefore, Jewish
day schools will probably have a good hit rate for their secular collection,
but not Judaica.
Many general source databases have minimal Judaica children's books
and materials, Jewish curricular materials, small press Jewish publications,
older and out-of-print Judaica, etc. If you want Judaica classifying
and cataloging (Elazar or Weine, or Judaic subject headings), then plan
to fill in or substitute your own call numbers and subject headings,
do original cataloging, or wait until other synagogue and Jewish day
school libraries put their library catalogs online.
••• Access for a fee
Lisa Handelman set up an account with MarciveWeb SELECT. She takes
a stack of books to her desk or works from home, and logs onto Marcive.
Lisa answers questions on screens as they walk her through the procedure.
Working with one book at a time, she keys in ISBN or title..., views
Marcive's full bibliographic record for that item, clicks OK, answers
more questions, adds notes, subjects, and other fields at no charge,
and, when finished, her book stack is represented as a list on the Marcive
screen, which she OKs and then logs off. In the morning, Lisa logs on
again and her file of MARC records is waiting for her to download and
import into her automation software program. Marcive bills her 27 cents
per recor d. If she orders label sets (8 cents), bar codes (5 cents),
and authority records (11 cents), they are delivered UPS in 2-5 days.
Lisa says that it is very easy, and requires no knowledge of library
standards because the Marcive program takes care of everything.
Vendor database charges vary—approximate costs appear to range from
25 cents to $1.50 per record, depending on the services, not including
start-up fees. If it is a large shared database, then there are usually
membership fees since cooper ative cataloging is one of its goals. Some
vendors charge for every search, whether or not successful (e.g. RLIN);
some for access, search, and downloaded record; others only for the
downloaded record. These databases have full MARC records and contain
the basic Library of Congress MARC records, among others. Be certain
to check customer support before signing on to any fee-based service.
In fee-based services, take advantage of their free hit rate tests
and cost estimate offers. Send them a sampling of 50-100 titles, and
they will do a free hit rate test to see how many of your titles are
in their databases. Be sure to include some unusual books and materials.
Compare their hit rate with others.
Marcia Goldberg shared an inexpensive way that they accessed OCLC a
few years ago: "...hired a student, and had a special account with OCLC
to do retrospective conversion. If you work after 5:00 PM weekly, and
on Sundays after 12:00 noon, it is half price, not including labor.
If lucky, you can do 60-70 books at a time."
Some vendors sell their databases on CD-ROMs because it is less expensive
than online access. CD-ROMs hold a couple of million cataloging records—to
search, tag, download, edit, and upload into your automation software.
Updates are usually monthly or quarterly. Susan Dubin used Brodart's
Precision One Cataloging System, which includes both print and non-print
MARC records of titles commonly held by school and public libraries,
including Library of Congress records. She is at a da y school and synagogue
library; her hit rate for the secular collection was excellent, but
not the Judaica.
For a list of vendors, ask colleagues for recommendations or refer
to sources at the end of this article.
••• Access for free
Free access means that you go onto the Internet and access Web-based
library catalogs; many allow you to view their full MARC records. This
allows you to do free copy-cataloging—either print from the display
and key the record data into your own automation softw are, or, if you
have the correct software, download the record, edit it for local needs,
and then upload it into your automation software. Recon and normal cataloging
can be handled the same way. There are numerous libraries available
that our Judaica librarians consider useful. Steve Bergson writes: "The
libraries we have used are University of Toronto, Harvard, Brandeis,
Toronto Public Library (North Division), Jewish Theological Seminary,
United States Holocaust Memorial Museum, Library of Congress, and the
University of Guelph; at present, most of the web-based library catalogs
that use Voyager software allow for easy downloading of catalog records
in MARC format (which need to be saved as txt files). Chaya Wiesmansearches
these libraries: Cleveland Public Library, Yale, Brandeis, Boston Consortium,
and any library on the Internet." Susan Dubinuses the Library of Congress;
but I've heard from others that it is often busy. Other sources for
MARC records as mentioned earlier are CASPR 's freeMARC, and CASPR's
LibraryCom where hundreds of libraries have their MARC records online
for free.
[An added note: Be aware that taking advantage of others' cataloging
work, for free, is frowned upon by some.]
• Consultant services
Sometimes, it is worth hiring a consultant to guide a library through
the recon process because there are so many aspects to consider and
so many decisions to make. When calculating comparative costs, it might
be a trade off—the fee for a knowledgeable consu ltant who knows the
process and has the prices at his/her fingertips vs. the salary of an
uninformed librarian required to do time-consuming tasks, e.g. research,
interview prospective vendors, develop criteria and study software packages,
execute the project.
Chaya Wiesman said: "[We] hired a library consultant of Library Automation
Inc. to walk us through each step, from choosing software to buying
bar codes and scanners, and getting best prices, to helping with technical
problems, and holding h ands when needed." Esther Nussbaum wrote: "Thanks
to our library automation consultant, we are proceeding at a rapid pace
with our records."
Besides those mentioned in this article, a good source for consultants
are your local and state, public and private, professional organizations.
Working with professionals who know the local scene and resources and
can share information, and mistakes, can be very helpful and encouraging.
• Combination of services
When planning the recon method, t hink carefully about your collection
and understand that no one method need be used for the entire collection.
You can mix and match, phase-in, fill-in, or whatever, as it suits your
content, funding, time schedule, personnel availability, etc.
Set up system for error detection, conflicting records, and authority
control.
Some errors can be corrected in a simple way. Florette Brill: "We have
had a problem with retro conversion. My advice to newcomers is to save
your shelf list, don't be too anxious to eliminate, you may need it
for reference purposes longer than you thought necessary."
An old adage in library work is that a mis-shelved book is a lost book.
Librarians often become frantic when they can't find a book that should
be there or upset when they find a book out of place. Ironically, many
of these same librarians are far more accepting when unable to find
an online record and or when incomplete subject headings skew their
search results. The same standards for the physical prope rty should
apply to electronic records, but they do not. We seem to think that
computer glitches are beyond our control, forgetting that most glitches
are caused by human error.
If integrity of your catalog is important, then detecting errors should
be of prime importance. This means you should set up a system for error
detection and correcting conflicting records. Some of this was handled
under "Establish cleanup method."
Authority control is essential, but is often ignored, resulting in
poorer retri eval, unless planned for during recon. If you wait until
the end when funding is gone and people are tired, a "we'll do it sometime"
attitude usually prevails. Authority control provides better access
to your collection because of consistency in names, subject headings,
spelling, punctuation, spacing, and other record elements. It also corrects
constantly changing subject headings and cross references.
Further, Amy Wissoker Graham wrote, "...authority control is matching
all of the name, subject, and series headings against LC's headings.
... Therefore, if a library has painstakingly changed every B.C. date
to B.C.E., or changed 'Jesus Christ' to 'Jesus of Nazareth,' these will
be undone by authority control." Other examples, "Holocaust--1933-1945"
will be changed to LC's "Holocaust, Jewish (1939-1945)," and LC Bible
headings are more complex too. This means that before you say "yes"
to recon vendors automatically changing your access points through an
authority control update (a common add-on produ ct of recon), you specify
those headings you don't want touched.
If you don't have the time or expertise to bring your online catalog
up to high quality standards, then try to get funding in your next phase
and hire a recon company, or companies that specialize in it. An example
is Library Technologies, Inc., whose specialized services include updating
MARC standards in the records, fixing filing indicators, building item
fields, duplicate record resolution, creating barcodes, and performing
authority co ntrol. It will make a big difference in your patron search
retrieval rate if quality control is truly under your control.
Choose whether to close or freeze card catalog while recon in process
"What do we do about the card catalog while recon is in progress?"
is a common and good question. One option is to pull the cards being
replaced by the online catalog, as recon is phased in—a time-consuming
task and worthwhile only if you have a reliable volunteer. Alternatively,
freeze the card catalog—th is means just leave it there—add no cards,
pull no cards, and correct no cards. The phased-in online catalog then
begins to duplicate the card catalog as recon progresses—a frustrating
experience while searching, but one patrons will have to learn to live
with. This is the usual choice for libraries. Be sure to post a note
on your card catalog to indicate that the card catalog is frozen and
that a new online catalog is being developed.
• Set realistic time schedule
Helene Tuchman writes: "For recon, we updated all our shelf-list
cards with ISBN or LCCN or good bibliographic data. . . . It took us
two years to prepare the shelf-list, send it out, load records, put
barcodes in the books and go online for circulation. About 1 year after
we began, we did use the system for an online catalog after the first
1500 records came back. We automated a collection of about 4000 items."
Our assumptions about time schedules are often under-estimated in library
work because the work is so detail ed and requires so much thought.
Then too, staff shortages, money limitations, and other reasons keep
us from meeting a deadline. Be sure to set a realistic time frame—one
within which you can live comfortably.
• Staff and patron training
Bottom line - if users don't know how to use it properly, much of
your work will be for naught. One of the oft-repeated comments by my
colleagues and myself during our graduate library school years was how
much helpful information was on a catalog card, which we had not realized
even in college. Share with your staff and patrons some of your recon
decisions, and teach them to read a record, so they can appreciate the
value of an online cataloging record.
In conclusion, the recon process is complex but very doable. Explore
your options, including many not covered here, because of space limitations.
As you plan, think of the many who have completed this process—and felt
satisfaction with the results. It can happen to you too!
Any questions, contact Barbara Leff at byleff@aol.com
Special thanks for advance critiquing of this article:
Andrew Wertheimer, AJL Automation Committee chair; research assistant
professor, University of Nebraska Libraries, Lincoln.
Amy Wissoker Graham, authority control editor, Library Technologies,
Inc., Willow Grove, PA
Sincere appreciation to AJL members who shared their experiences:
Susan Berson, Congregation Emanuel, Denver, CO
Steve Bergson, Jewish Public Library of Toronto, Toronto, Ontario,
Canada
Florette Brill, Neipris Library, Temple Israel, Sharon, MA
Rick Burke, University of Judaism Library, Los Angeles, CA
Susan Dubin, Valley Beth Shalom Synagogue Library, Encino, CA
Marcia Goldberg, Gratz College, Melrose Park, PA
Amy Wissoker Graham, Congregation Ohev Shalom, Wallingford, PA
Lisa Handelman, Adat Ari-El Synagogue/Day School Library, North
Hollywood, CA
Adaire Klein, Simon Wiesenthal Center Library/Archives, Los Angeles,
CA
Esther Nussbaum, Ramaz Upper School, New York, NY
Helene Tuchman, Temple Emanuel of Newton, Newton, MA
Chaya Wiesman, Ramaz Upper School, New York, NY
Edythe Wolf, Omaha Jewish Federation, Omaha, NE
Abigail Yasgur, Jewish Community Library of Los Angeles, Los Angeles,
CA
AJL-NYMA (AJL-New York Metropolitan Area) Chapter- February 25, 1999
- Panel on Retrospective Conversion and its Application in Judaica
Libraries.
Sources mentioned in this article as examples:
Brodart Co., 500 Arch St.,
Williamsport, PA 17705. 800-233-8467, X581. For Info: salesmkt@brodart.com
Precision One Cataloging System - for both recon and on-going cataloging
CASPR Library Systems, Inc., 100
Park Center Plaza, Suite 550, San Jose, CA 95113. 800-852-2777
LibraryCom (Library Directory
& Hosting Service) For Info: webmaster@librarycom.com
FreeMARC
Library Conversions Limited Company, 3746 East Cholla St., Phoenix,
AR 85028. 800-268-9278. For info: libconv@primenet.com
Marcive, P. O. Box 47508, San
Antonio, TX 78265. 800-531-7678 For info: info@marcive.com
OCLC Online Computer Library Center,
Inc., 6565 Frantz Road, Dublin, OH 43017. 800-848-5878
Booklet: OCLC
- Retrospective Conversion - Guidelines for Libraries
Some Sources for Products and Services
The Librarian's Yellow Pages 1999. Larchmont, NY: Garance, Inc.
(352pp.)
Free book - write on your letterhead to Librarian's
Yellow Pages, P. O. Box 179, Larchmont, NY 10538. 800-235-97 23.
For info: info@LibrariansYellowPages.com
Library of Congress A
List of Commercial Suppliers (1/98)—as an alternative to SelectMARC
Retrospective Conversion Services [a discontinued service]
Librarian's Online Warehouse
LibraryLand
Software products mentioned in this article as examples
MITINET/marc: Information Transform,
Inc., 502 Leonard St., Madison, WI 53711. 800-tag-marc. info@mitinet.co
BookWhere? for Windows 95/NT:
Sea Change Corporation, 66 95 Millcreek Drive, #1, Mississauga, ON L5N
5R8, Canada. 800-661-7274.
eZcat for Windows and PowerMac:
Book Systems, Inc. 721 Clinton Ave., Suite 7, Huntsville, AL 35801.
800-219-6571.
Consultants mentioned in this article
Library Automation Management,
Inc., 101 Clark Street, 27C, Brooklyn Heights, NY 11201. 718-834-1414
For info: webmail@LibraryAutomation.com Contact: Robert
Rowen
Library Associates,
8845 West Olympic Blvd., Suites 100-200, Beverly Hills, CA 90211. 800-987-6794
For info: info@libraryassociates.com
Contact: Deborah Schwarz
Library Technologies, Inc., 2300
Computer Ave., Suite D-19, Willow Grove, PA 19090. 800-795-9504 For
info: LTI@libraryTech.com Contact:
James Schoenung, Ph.D.
Note: Opinions and recommendations in this article are those of the
author and the people cited only, and do not represent an endorsement
by the Association for Jewish Libraries.
|
|