|
Administrative
|
Technical Services |
Cataloging Workflows Task Force
Final Report
Dec. 8, 2004
Members: Becky Atack, Alice Cunningham, Bryan Kurowski, Monica Sanchez, Pete Wilson (chair)
Primary
reporter: Pete Wilson
This
report has been a long time in coming.
The Task Force was formed in December 2002 to investigate what changes
in cataloging organization and workflows might enhance efficiency. We worked through July of 2003, then went
into a sort of dormant state because the chair, Pete, found his time squeezed
hard by family responsibilities for the rest of the year and a little
longer. Some time after these eased up,
the Task Force resumed its work (helped by a little prodding from Roberta), and
set a deadline of December 1, 2004, for turning in the report. And we very nearly made the deadline!
None
of the changes that we suggest in the report (or have already had implemented)
are massive. The physical facts of
staff being in two widely separated locations, as we saw it, restricted
possibilities for personnel reorganization, as do the realities of how many
different kinds of work can be asked of Order Services staff members at their
current job classifications. Another
boundary for our considerations was that we considered almost exclusively
monograph cataloging, not serial cataloging.
That seemed to be where the issues were, and certainly serial cataloging
is a pretty small fraction of the entire cataloging operation. However, we’ve made quite a few
recommendations, large and small, about cataloging procedures, many of which
have already been implemented. And the
report gives attention to rather broad issues like training, documentation, and
questions about the inventory and the INVD process.
Thanks
to all the members of the Task Force for participating in a process that took
longer than it should have, largely because of its chair.
1. Staff organization
We recommend that the division be hesitant to ask Order Services staff to take on new cataloging duties that would require more attention and revision than they currently give the straightforward LC copy that they catalog. There are several reasons for this:
a. Acquisitions work is complex and demanding enough in itself.
b. Taking on more difficult cataloging duties would be difficult without an expert cataloger at Baker to handle training and questions (this could be remedied, but probably at the cost of disruption in other areas—and the solution might still be unsatisfactory, given that a single cataloger at Baker without access to other catalogers for consultation or to our reference materials would be pretty handicapped).
c. It would probably be unrealistic and unfair to expect Order Services staff to learn a whole new set of skills and take on a whole new set of duties without increases in pay and higher job classification, which seem unlikely at this time.
There has actually been one notable addition to OS cataloging duties since this task force was formed—two people, Alice and Gina, are “finishing” CIP copy according to a procedure drawn up by Becky. There are no plans currently to have other OS staff members do this. The Task Force had nothing to do with starting this but it is good for efficiency in that it eliminates “handoff” for items which need very minimal cataloging input.
It is at least theoretically conceivable that full-level PCC copy could be handled in speed cataloging just as LC copy is now. Access points on PCC copy are supposed to be backed up with authority records just like LC copy. In practice, though, making this change would represent a considerable concession in terms of record quality—the catalogers have found PCC records to be more unreliable than LC records, even though the LC records have slipped in quality somewhat themselves. Thus, this change should not be made without some study of the quality of available PCC records, something we didn’t do in the Task Force.
One idea discussed in Task Force meetings was to have OS staff do INVD processing for items which the original catalogers select for dispersal while reviewing their shelves in the Baker mail room. Currently these books are boxed up and sent to the cataloging team to be processed. (This idea came from one of the OS members of the Task Force, by the way.) It does seem that training a staff member at Baker to process books selected for INVD and then send them straight to Preservation might enhance efficiency. (Decisions would have to be made regarding authority work and probably other issues.) When asked about this idea at a weekly meeting, however, catalogers were not enthusiastic. Farming out INVD books to OS staff would represent a certain loss of discretion over their handling that appeared to be unwelcome, and there was concern over authorities issues. (A later section of this report will discuss inventory dispersal further.)
The Task Force discussed some longer-range possibilities for redistribution of cataloging duties. It seems a good possibility that at some point the cataloging unit will be ejected from the library to make more space for books and users. While this won’t please the catalogers at all, it might allow for the reunion of cataloging and acquisitions staff in one place. This would mean that OS staff members who took on expanded cataloging duties could get day-to-day support much more easily and frequently from other catalogers. They could also take part in the daily passing-back-and-forth that the original and copy catalogers currently do to get a book where it needs to be for accurate and efficient cataloging. Initial training would also be much easier and less disruptive. Of course, as we said above, there would be issues of wages and job classifications—but here we are talking long-term possibilities, not current plans.
Proposed changes are either addressed
elsewhere in report or postponed for now.
2. Re-searching
To
whatever extent it actually appears to be cost-effective, re-searching of
cataloging copy should be done at Baker before items are routed to
catalogers. The Task Force made certain
recommendations last year which have already been made policy.
Currently
these categories are re-searched:
a. All items received on provisional (“NC”)
records.
b. All rush items handled in OS which do not
have LC copy on Acorn.
c. Items which upon receipt turn out to be a
different edition from that actually ordered.
We
did an experiment on re-searching books which had CIP copy and found that that
didn’t seem particularly fruitful.
(Since then, Alice and Gina have begun cataloging—not re-searching, but
fully cataloging—CIP copy at Baker; see above.)
Re-searching
has resulted in many items being speed cataloged at Baker which previously
would have gone to catalogers first, and has also meant that items can be
routed to copy catalogers rather than to original catalogers. It also means that more up-to-date records
are in Acorn for items awaiting cataloging, which should be of value to the
users in that the records can be more easily found and also will better
describe the items. We feel that
re-searching has been and will continue to be beneficial in ways that outweigh
the increased time spent on these items in OS.
To
ensure optimum efficiency, catalogers who receive items which have been
recently re-searched at Baker should avoid “re-re-searching them,” to save OCLC
searching costs and prevent redundant work, unless the items sit uncataloged
for some time. The task force came up
with codes for OS staff to add to re-searched records so that catalogers would
know when the re-searching was done.
Re-searching
turned out to give rise to one unforeseen problem for which we think we have a
passable solution. When OS staff began
to overlay previously-“NC” records with LC copy and speed catalog the items,
name and subject headings were brought into the system without triggering our
receipt of corresponding authority records from Marcive, since records brought
in by overlay do not automatically go to Marcive for processing. This created additional work for authorities
staff. This work can be avoided if OS
staff add a “954 code” (cataloging team members add these routinely in certain
situations already) to overlaid bib records after speed cataloging. This code (it goes in a 954 field and
consists of the month and year, e.g., 122004) triggers the system to send the
record to Marcive for processing, which gets us the matching name and subject
authority records we need. It would be
most efficient to add this 954 field only when there are unauthorized headings
in the record; if there are no unauthorized headings there is no need for
it. However, Mary Charles thinks that
it would be more efficient in terms of time and money to simply add a 954 to every
record that has been overlaid and then speed cataloged, than it would be to
never add them. Thus, if it is a
stumbling block for OS staff to check for unauthorized headings in these
situations, we recommend that they add 954s without checking.
TechForce
recommends addition of 954 to every record that has been overlaid and speed
cataloged.
3. Recommendations for training of paraprofessional
catalogers
Training of new staff in Cataloging has traditionally been done primarily in-house and on an individual basis. In general, outside training is not cost effective, except in cases involving large numbers of staff. Vanderbilt is fortunate to have two catalogers, Ann Ercelawn and Mary Charles Lasater, who are recognized nationally for their expertise and teaching abilities in the fields of serials cataloging and authorities control, respectively. Staff who have had the privilege to be trained by them have had the best possible instruction. Most of the other original catalogers contribute to staff training either directly, as does Don, the primary instructor for original cataloging, or more indirectly through their supervision and consultation of support staff. All of the professional catalogers are unfailingly helpful to support staff and to each other. It sometimes “takes a village” to catalog!
Vanderbilt is unusual in that it has a high percentage of original (i.e. professional) catalogers who actually catalog and who work with member copy in many formats. However, the trend nationally seems to be toward greater responsibilities for support staff, and there is no reason to expect this will not be true for Vanderbilt. For example, in the libraries at UT-Knoxville, the cataloging of all books using any available copy is done almost entirely by support staff. To meet this challenge, support staff have undergone intensive training (now in-house, but initially by an outside trainer) by the professional catalogers, who spend very little of their own time actually cataloging. As a result of this greater responsibility being given to support staff, they have received upgrades in job classification and corresponding pay increases (something we would see as essential to similar changes in our future).
One
of the virtues of UT’s training is its intensity, involving frequent and quick
review of the trainees’ work. Unfortunately, since Vanderbilt’s professional
catalogers have so many responsibilities in addition to their own cataloging,
at times the training of support staff has been hindered. Another problem
hindering the timely review of trainees’ work is the physical separation of Technical Services teams. There
are also other considerations to keep in mind when comparing the cataloging
staff at Vanderbilt to that of UT.
Since UT is not a BIBCO contributor, their cataloging staff members have
not required that additional training. For the most part they also lack
language and subject expertise. When staff have these skills and knowledge,
they often see the need for cataloging that is much more detailed, and thus
more time-consuming.
Good
training, well organized and well thought-out, is essential for all staff,
especially in light of both the declining quality of bibliographic records
nationally and the constant change in staff roles and responsibilities. In particular, the decline in the quality of
national-level bibliographic records has resulted in an increase in the amount
of time spent on improving these records, whereas in the past the quality of
copy used by copy catalogers was much more reliable, and the records required
less modification.
From
our readings and discussions with colleagues, we have found that what is
lacking at Vanderbilt, despite the wealth of talent, is a coherent training
curriculum that builds on acquired knowledge, that is based on well-defined
concepts and policies, and that can be adapted to the individual needs and
levels of trainees. Catalogers need to review and update policies and formulate
a basic curriculum that includes the following:
a.
instruction in the use of the wide variety of necessary work tools, both print
and online.
b.
definitions of the levels and types of cataloging in regard to local practice.
c.
instruction in the cataloging of print materials.
d.
instruction in the cataloging of non-book materials, including electronic
resources.
e.
timely completion of a series of basic cataloger training sessions that are
focused on specific areas of cataloging protocol (e.g. NACO and LCSH training).
f.
clearly documented procedures (see below).
The TechForce
recommends a new CAT Training Task Force to look at both cataloging
training and documentation. Some members of the CAT Workflow Task Force
have already volunteered to serve.
4. Recommendations for improving and
maintaining documentation
Thanks to the time and efforts of Ann Ercelawn and Mary Ellen Wilson, the Technical Services web page has updated documentation that is easy to access, for the most part. What is missing, however, is an online manual for cataloging. We were deeply impressed by the quality of some of the manuals we viewed online, especially those which had a wealth of easily accessible information, clear and well-designed formats, and, best of all, procedures that were very similar to Vanderbilt’s. The best of those manuals, while still a work in progress (as is everything related to technical services these days) is the University of Virginia’s:
http://www.lib.virginia.edu/cataloging/manual/
Most fortunately, Virginia’s library uses Sirsi. A training manual based on the model of Virginia’s training documentation would prove to be the quickest and least labor-intensive approach to creating the necessary training resources here at Vanderbilt.
Also, we believe it would be extremely helpful and efficient to have an online training document in the form of a MARC record template with hypertext links to documentation for each field and to create “cheat sheet” reference documents that highlight crucial points, procedural steps, etc. Both of these would save time for staff.
See #3
5. Inventory and inventory dispersal issues
Last
year the Task Force proposed that the inventory be “frozen”—i.e., no more books
should be added to it. This change was
adopted by the cataloging team on a temporary basis, but the time set for
reconsideration of the policy has long passed without serious discussion.
The TechForce
recommends that the Inventory remain frozen.
When
this change was made, and for many years before, the inventory was composed
mostly of books which were given very low priority for cataloging. Items were frequently added to it but it was
no longer the active cataloging backlog that it was when large numbers of items
were sent to it routinely and bibliographers regularly selected items from it
for cataloging. Freezing the inventory
was in part a recognition of this change in the way it is perceived. It also served, especially in conjunction
with the work of the “R.I.P.” task force (which was much more instrumental in
this), to help reduce the inventory and thus Technical Services’ costs. Inevitably, of course, this change meant
that some low-priority books would simply sit longer in the Technical Services
workroom, but this does not appear to have been very burdensome. A shelf is maintained in the workroom for
items which have been examined by catalogers and found too marginal, and too
difficult to catalog, to be worth spending much time on at the moment, but this
single shelf has never been more than half full, and at this writing it has 11
books and 1 microfilm item on it.
Putting a book on this shelf is understood as an invitation to any other
cataloger to tackle it.
Inventory
dispersal is, of course, a related topic.
(Here we’ll discuss the processing of newly received books for
dispersal, not the efforts to pull books from the existing inventory and get
them on the shelves.) When the Task
Force discussed the idea of having OS staff process books selected at Baker for
INVD (see above), we were struck by the fact that the purposes of INVD are
(have become?) somewhat unclear. Is it
a mechanism for immediately sending to the stacks books considered of minimal
importance? Or is it a holding bin
where books await re-searching at a later date (much like the shelf described
in the last paragraph) until we finally give up? What makes a book appropriate for INVD—marginal importance, or
difficulty of cataloging, or (conversely) a record that already seems
serviceable enough not to require review?
Do we need to get all headings authorized before INVD’ing a book
(something of a paradox, since authority work can often be the most
time-consuming part of cataloging)? The
Task Force did not think we could answer these questions ourselves; we would
recommend the cataloging team consider them.
Answering these question would help clarify whether having OS staff do
INVD processing would be a good idea, and it would also relate to efficiency
questions regarding how much time the cataloging team spends on marginal books
selected for INVD.
It
is probably appropriate to mention, by the way, that we asked John Haar whether
he is in favor of continuing inventory dispersal, and he said yes. His comments might be helpful in deciding
exactly what INVD should be:
… it gets books to the shelves
faster. I recall that prior to the
creation of the program, we
conducted a test and discovered that
inventoried books stored in the
Central stacks were used at a
higher rate than those stored in
the Annex.
Since these books tend to be
recent publications, I think placing
as many of them as possible in
the stacks is preferable to storing
them at the Annex, which
generally works better as a storage
space for older materials.
I appreciate that the catalog
records are sometimes incomplete,
but this would also be the case
if the materials were at the Annex,
so the real issue, if I
understand it, is only their location.
Obviously cataloging everything
quickly is the best option, but if
that's not possible, I think
inventory dispersal should be expanded.
From the last paragraph quoted, I conclude that John would rather get INVD books to the shelves fast than attempt to improve their records. While changes to the INVD process probably should await the cataloger discussion recommended above, we do suggest that books selected for INVD not be re-searched more than once before they are processed, since to retain them for very long defeats the purpose described by John.
One change, already implemented, that I believe was recommended last year by the Task Force was the placement of books selected for INVD on a shelf where all catalogers can easily look at them and pick them up for true cataloging if they want. If there is to be any “waiting” at all involved in their processing, it seems appropriate and helpful to do this.
TechForce response: Catalogers
have never embraced the INVD procedure. As a result, the INVD-selected items sit
on CAT shelves for a long time with the hope that someone will find time to
catalog them fully. This is not in the interest of patron access. TechForce
recommends replacing the INVD procedure with a new procedure that would allow for
minimal cataloging of marginal items. These would go to the shelves as finished, with no
intention of later cataloging upgrade. We would work with
CAT to define “marginal” items and “minimal”
cataloging before putting the procedure in place. We would
also need guidelines about the level of authority
work required on these items. INVD would
no longer be used.
6. Speeding up the processing of books that are urgently needed
Last year, the Task Force recommended a turnaround time in cataloging of three days for patron requests. Previously non-rush patron requests had not had firm deadlines, and special processing done in OS was being wasted when the books remained in cataloging for an excessive time. This policy was implemented but has since been superseded by “Level 1 priority” rules, which give even higher priority to all patron requests.
Another matter discussed by the Task Force but tackled elsewhere was the idea of circulating urgently needed books before cataloging when cataloging cannot be completed in an extremely brief time. We certainly endorse this change, since the ultimate point of cataloging is to get books to people who want to use them.
The TechForce recommends more frequent decision to
circulate uncataloged urgently needed books. We also
will investigate improvements in the
notifications to patrons when items are ready, since this currently depends on overnight Acorn system
reports.
7. Gift
processing
The Task Force proposed last year to allow verifiers to send gift items directly to catalogers when their initial processing required cataloging decisions beyond the scope of the verifier. A good example would be an item that might or might not be a second copy (or an added volume) on an existing record. A verifier’s incorrect decision in such a case usually causes inconvenience for the cataloger, and wastes both the verifier’s and the cataloger’s time, so in this case we are probably better off having a cataloger process the gift from scratch. Pete wrote a short procedure to help catalogers with coding Acorn records for these gifts. This change has already been instituted, but in reality it appears the verifiers have rarely simply sent gift items to catalogers without warning; they contact the catalogers about the problem beforehand.
The Task Force also recommended last year that “replacement” gift processing be moved back to OS. At the time it had become the responsibility of the cataloging team, but it appeared it would be more efficient to include these items with regular gifts. The change was made.
Done.
8. Whether to accept UKM records from PromptCat
The Task Force concluded that it appeared more efficient in terms of time and money to continue not to accept UKM (British Library) records from PromptCat, even though PromptCat has revised their record selection criteria so that they will not routinely select UKM records when superior records are available. See Appendix 1 for a report by Pete about a study he made on this.
OK—no change
9. “LC diff ed.” flags
A copy cataloger brought the declining use of “LC diff ed.” flags to the attention of the Task Force. These flags are inserted when an OS staff member finds that while there is only member copy for an item, there is LC copy for a different edition of the same item which can be used to simplify greatly the cataloging of the item in hand, thus saving time. The Task Force discussed the fact that such flags seem less common than they used to be. The OS members of the Task Force said that there has never been a decision to discontinue the use of these flags, but possibly Smartport searching and searching by ISBNs have made it less likely that a verifier will happen across one of these different-edition records. To search deliberately for other editions of the same title seems inefficient; that decision is probably better left to a cataloger. To change OS searching procedures in general to produce more “LC diff eds.” would be a disproportionate remedy, to say the least. The Task Force decided to recommend that all OS staff insert “LC diff ed.” flags in books when they do happen to find such a record for a different edition, but that they not look for other records exclusively for this purpose.
OK—no change
10. Changes to OS processing procedure documents
The Task Force recommended that the list of “approved libraries” (i.e., libraries whose cataloging copy, in particular subject areas, was deemed equivalent to LC copy) be removed from OS’s documentation. This wasn’t really about efficiency as much as it was an acknowledgement that the list was rarely used by OS and probably way out of date. The cataloging team agreed with this recommendation and the change was made.
Another change was made to OS’s “How to determine LC, CC, MC coding for books” document. This document includes a chart showing what combinations of source code, encoding level, 040 and 042 mean that an item can be coded as LC for routing purposes. It indicated that the source code had to be “d” for an lccopycat record to be treated as LC, but really such records can have source codes of “blank,” “c” or “d.” Also, the chart indicated that a Bibco record (i.e., one with “pcc” in the 042 field) had to have source code “c,” but many actually have source code “blank.” Zora changed the document with Mary Ellen’s approval.
Done
11. 050/090
division of copy between copy and original catalogers
The aforementioned “how to determine coding” document, in effect, bases the decision of whether to route a book to the LAIV copy catalogers (“CC”) or the original catalogers (“MC”) on whether the call number in a record is in an 050 or an 090 field. (Items without call numbers go to the original catalogers.)
This is actually an extremely arbitrary criterion. It is entirely up to any OCLC member library whether they record their call number in an 050 field (with proper indicators) or an 090 field. The catalogers’ general sense is that records with 050s do tend to be somewhat better and more complete than those with 090s, but in principle there is no distinction. (PCC records do routinely use 050s, but those are routed as “LC” in the first place, and don’t enter into this issue.) This means that non-LC copy is divided up between original and copy catalogers on a basis almost as arbitrary as, say, what color the binding is or whether the author is male or female.
This has been a bugaboo for the Task Force chair for years now. No one else is nearly as exercised over it as he is, but the rest of the Task Force does acknowledge the arbitrariness of the criterion. Neither Pete nor anyone else has been able to suggest a more theoretically grounded way of dividing the cataloging up, though. In practice it has not presented obvious problems, because the LAIV’s are very capable catalogers who can handle most of the member copy that is routed to them, and because day-to-day communication and proximity make hand-offs of material between them and the original catalogers extremely easy and fast. Furthermore, any more scientific way of dividing up the work would probably require more inspection by OS staff, and that might very well negate any positive impact. So we might as well keep doing what we are doing.
OK—no change for now
12. Thesis cataloging
We could reduce our work by cataloging Vanderbilt theses and dissertations at a more minimal level, as do many other libraries. We don’t recommend this, however. Pete asked Zora and Mary Charles about why we catalog theses fully, and was told that when Malcolm Getz was the library director, he was adamant that we consider theses just as important as other materials (or more so) and continue to give them full level cataloging even as other libraries were, for example, no longer making subject headings for them. Of course, Getz is long gone, but his position still seems valid. If we don’t fully catalog our theses, nobody else will, and it seems part of our responsibility to the University to make its own scholarship as accessible as possible. Efficiency means doing the work that is needed, not just working fast, so we don’t think downgrading our cataloging of theses would enhance efficiency. We mention the issue here just to answer in advance any challenges that might arise in the future.
No change for now
13. Notifying
catalogers when online resources need to be cataloged
This issue came up in discussion during the next-to-last Task Force meeting, when a deadline had already been set (and was rapidly approaching) for completing the report. Also, it is a rather complicated issue absolutely requiring information and cooperation from staff outside Technical Services. Thus our work on it, and our recommendations in regard to it, are quite preliminary in nature.
A while back, the catalogers had to spend a considerable amount of time on a special project. The Central Library had had a student worker go through the Web electronic database listing to check whether each entry was represented in Acorn. According to Zora, there were about 140 titles that were problems. Although Central did the project, the problem titles were associated with other libraries as well—Management, Science, Law, Biomedical, etc. Many titles were not represented in Acorn. In some cases there was an Acorn record associated with the title, but it was for a previous or related title or the URL had been changed on the database list but not in Acorn. If I remember correctly, for some titles Acorn records existed but had not been touched by a cataloger—thus, the title was ordered on Acorn and successfully activated, but the cataloging was never completed. (That situation has been noticed many times independently of this particular project.)
This experience emphasized what we had already realized—that sometimes online resources which are added to the Web database listing (i.e., the DB/Text database) fall through cracks and do not get fully cataloged in Acorn, at least for a significant period of time. (Here, by the way, we are speaking of online resources which are not E-journals. E-journals carry their own complications—in particular, that we get records for most, and in the future perhaps all, through SFX—but some of the recommendations in this section may be germane to them to some degree.) Many online resources are ordered through Acorn, on provisional or OCLC member copy records, but these are often inadequate and usually have no URL or an incorrect URL—if a cataloger is not alerted to the fact that the resources are now available and listed in the database listing, the records will remain inadequate. Clearly, this is a question of cataloging efficiency, since not doing the job is always inefficient! It also appears that when a URL changes or some other change in a resource occurs, while the change may be noted by a bibliographer and the Web listing updated, the news may not always get conveyed to a cataloger to update the Acorn record. Additionally, there are free online resources which are not ordered, but merely added to the Web listing because a bibliographer finds them useful. The bibliographers should send requests to the cataloging team via a Web form to have these cataloged in Acorn, but it is uncertain whether this always occurs. Even if this generally happens as it is supposed to, it might be useful to make cataloging follow-up more automatic—and efficient—when a resource is added to the Web listing.
Finally, according to Chris Waldrop, who was very helpful on this issue, we sometimes order online resources through Acorn, but are not notified by the publisher when they have processed our payment and activated our access to the resource. Clearly, since an appropriate bibliographer or administrator at the ordering library is not notified, the Web listing is not updated to include the resource; nor can anyone alert a cataloger to its availability. Thus the Acorn record for the resource continues to have uncataloged status and to be inadequate for our users. This is more than a cataloging problem, of course, but a solution to the cataloging issue will have other benefits as well.
Pete, Chris, Zora, Richard Stringer-Hye and Mary Beth Blalock discussed these issues in an exchange of emails over a few days. Little was resolved at that point but two ideas were proposed which may remedy the problems we see. Given the deadline for this report, the Task Force doesn’t have time to specify details of these solutions or even necessarily vouch for their benefits, but we do suggest they be considered for implementation:
a. Richard suggested that a script might be written so that a notification email would be generated automatically when a resource is added to the DB/Text database for the Web listing. (The same thing might be done for the Web e-journal database, but the follow-up would likely be different.) The notification email would certainly go to someone in Cataloging and Authorities; it could perhaps go to a group mailbox (Zora’s idea). In any event the email would be forwarded to the appropriate cataloger, who would fully catalog the resource in Acorn. This would cover both purchased databases, which would already have preliminary records in Acorn, and those free databases added to the Web listing by bibliographer choice, which would not. Zora suggested that a similar notification could be generated when a record in the DB/Text database was changed—to update a URL or change a title, for example.
Currently we rely on a bibliographer to take the initiative to fill out a request form to get an online resource cataloged in Acorn, on top of the work done to get it into the Web listing; this is an opportunity for things to fall through the cracks. The proposed change plugs those cracks. A little investigation might show us how big those cracks really are—i.e., whether or not “staff vigilance” (in Rick’s words) is already doing the job for the most part. (Rick does caution that we may make changes in the way the Web listing is created in the near future, and thus his idea may have only short-term usefulness—that will have to be considered.)
b. Chris Waldrop offered to work up an addition to OS procedures so that claim notices are generated by Acorn when an order for an online resource goes apparently unfulfilled. He suggested a weekly “show to” claiming mechanism could be used for this; I presume the claim would not be automatically sent to the publisher. (Chris can investigate whether a check has been cashed, which will of course be valuable information in trying to determine the status of an online resource order.) Apparently claiming for online resource orders is currently not done or is inadequate. With this change Technical Services would, as Chris says, take responsibility for checking on the orders, and would not have to rely on a bibliographer’s complaint to find out about a delay. Beyond brief communication with Chris, the Task Force hasn’t explored this idea and we are not at all clear on the details of how it might be implemented, but we like it in the abstract. It seems a likely route to improvement in getting online resources available through Acorn in a timely fashion. It should also get them in the Web listing faster as well.
TechForce
hopes
that an Electronic Resource Management System will be available soon. Until
then, we will carry on with existing procedures,
but try to follow them more closely. We will
investigate the feasibility of claiming online resource orders.
14. Outsourcing
The Task Force agreed that outsourcing was somewhat tangential to our charge. It certainly can be useful, though, when material to be cataloged is beyond the capabilities of our catalogers—for example, music items back when Blair lacked a cataloger and Jeff had not developed his music cataloging skills, or material in very obscure languages or non-Roman scripts. When the material in question is within our catalogers’ reach, though, we advise being very cautious about outsourcing projects. Our experiences with TechPro and the Sigaux collection have proven that a lot more local staff time can be eaten up than one would expect, and on the other hand I believe our local handling of the first big Helguera batch was fairly successful. In regard to any future outsourcing project, we should review both the problems of one and the success of the other, and planning should involve staff members who have had to deal with cleaning up the results of past outsourcing projects.
TechForce agrees.
15. Recommendation for continuing attention to cataloging efficiency
Our report is being submitted just a few months before we will begin using OCLC’s Connexion interface for cataloging. This is just one example of the constant changes we undergo in the mechanics of our daily work. The still-increasing importance of electronic resources, online or not, is an example of how the changing academic and information environment forces us to re-examine needs and priorities. We may have to deal with sudden and massive changes as well—a move out of the library, as mentioned in section 1 above, would create many inconveniences for catalogers but might facilitate reorganization of cataloging personnel.
These examples illustrate why the Task Force’s work on matters of cataloging efficiency (which certainly none of us consider definitive as it is) must be continued and extended in the future. There is probably not enough need to maintain a standing committee purely devoted to cataloging workflows, and we members of the Task Force specifically are not volunteering to become one. Rather, Tech Force and the Cataloging and Authorities Team should stay vigilant about whether changes in our software, equipment, personnel, and physical situation, or changes in what types of materials the libraries need and their relative priorities, make established procedures less efficient than they used to be. Bigger changes will require the most imagination, and probably task forces will need to be appointed to wrestle with them. But Tech Force and the Cataloging and Authorities Team can take the responsibility for anticipating and identifying these challenges.
TechForce agrees!
Accompanying the report are two appendices. One is a report on the rather minimal results of our survey of other libraries on their cataloging workflows. The other is Pete’s study on the question of whether to begin accepting UKM records from Promptcat again.
(This is the final version of the report, incorporating changes discussed at the Task Force’s Dec. 8, 2004 meeting.)
Appendix 1 to the Cataloging Workflows Task Force report.
STUDY
ON WHETHER TO RESUME ACCEPTING UKM RECORDS FROM PROMPTCAT, by Pete Wilson
(written Oct. 2004).
In
trying to figure out whether we should start receiving UKM records from PromptCat,
I looked at 6 PromptCat reports, all quite recent--for some of them, some of
the books had not yet been processed at Baker, but others had, so I figured all
the reports could be considered either very recently finished or in process.
1. Non-UKM
records were available as of processing time for 46% of the books that
PromptCat had a UKM record for (48% when adjusted—see below).
Here
are the reports, plus how many titles on each had UKM records, which of course
we did not receive:
bwk.040912 52
UKM records
bwk.040918 12 UKM records
bwx.040926 7 UKM records
bwx.041003 10 UKM records
ydx.040921 1 UKM record
ydx.040922 2 UKM records.
Many
of the titles were still represented only by a UKM record at the time they were
processed by verification staff (or there was still only a UKM record when I
checked OCLC, in the case of items not yet processed by verification
staff). Here’s a rundown on that:
bwk.040912 27 of 52 titles still had only a UKM
record available
bwk.040918 10 of 12 titles still had only a UKM
record available
bwx.040926 2 of
7 titles still had only a UKM record available
bwx.041003 4 of 10 titles still had only a UKM
record available
ydx.040921 1 of
1 title still had only a UKM
record available
ydx.040922 0 of
2 titles still had only a UKM record available.
The
remaining titles (except 2 in bwk.040912, which were unusual cases that I’m
disregarding) had non-UKM records available at the time of processing (or my
checking on OCLC):
bwk.040912 23 of 52 titles had non-UKM records
bwk.040918 2 of 12 titles had non-UKM records
bwx.040926 5 of
7 titles had non-UKM records
bwx.041003 6 of 10 titles had non-UKM records
ydx.040921 0 of
1 title had non-UKM records
ydx.040922 2
of 2 titles had non-UKM records.
Here’s
the same information in more summarized form:
Blackwell
UK: 37 (60%) had only UKM; 25 (40%) had
non-UKM records.
Blackwell
US: 6 (35%) had only UKM; 11 (65%) had
non-UKM records.
Yankee: 1 (33%) had only UKM; 2 (67%) had non-UKM
records.
All
books: 44 (54%) had only UKM; 38 (46%)
had non-UKM.
Obviously
the tendencies of Blackwell UK books are slightly different from those of
Blackwell US books. It would be good to
know in what proportions we really get UKM titles from each vendor—-the data so
far doesn’t really tell us that. In
order to get some idea of these proportions, I added up the UKM books from each
vendor for April, Aug. and Sept. 2004.
It turns out that during those periods, we got 180 UKM books from
Blackwell UK, 126 from Blackwell US, and 38 from Yankee. Applying the percentages quoted in the last
table to those proportions, we find that, theoretically, 48% of the books for
which PromptCat has UKM records will actually have non-UKM records at the point
of verifier processing.
B. The non-UKM
records almost never are appropriate for speed cataloging or LC copy
cataloging, and rarely even for routing as CC.
But they are more useful than the corresponding UKM records.
Of the 38 cases where a non-UKM record was (or will be) available at the point of verifier processing, a DLC record was available for only 1. (It was not actually brought in by the verifier—-one of two UKM records was used instead—-but I think it is probably appropriate for the book.) In 7 cases, the non-UKM record has an 050, and thus would be routed as CC rather than MC. In 15 cases, the non-UKM record does have a call number in 090 and would be routed as MC. In the remaining 15 cases, the non-UKM record has no call number at all and would also be routed as MC. 2 of the MC records with 090’s and 10 of those without have an encoding level lower than K (7 or M). A lot of these are M-level records from EQO (Univ. of Oxford), incidentally.
On the other hand, the UKM records PromptCat would be sending us for these titles are pretty useless. I ran through them rather quickly, but I believe in 36 of 38 cases they have encoding level 8 and no call number. Catalogers know that such records often include incorrect name headings and are totally unreliable for descriptive data. (One record had a blank encoding level but no call number; the other still had encoding level 8 but otherwise had been upgraded considerably by another library; I can’t be sure whether it had been upgraded by the time PromptCat pulled it from the database for use, but it seems likely.) I should mention that I am describing what these UKM records look like now in OCLC—I’m not reconstructing what they looked like when PromptCat chose them to represent the books. These UKM records get upgraded slowly, if at all—they are sort of like red-headed stepchildren, and the attention goes to the other records for the same titles in OCLC. Presumably many of these UKM records are merged into other records eventually.
Generally, these UKM records have very few holding libraries—rarely more than 1 to 4. Other libraries hardly ever use these records for their cataloging. We could use these records, of course, and update on them in OCLC—probably the record would eventually be merged with whatever other record might exist (usually with more holding libraries on it). But at least for a while, our holding symbol would be on a sort of orphan record, not on the record most libraries are using. And it would probably be much easier for the cataloger to overlay a better record than to dress up a UKM record until it’s presentable, to say nothing of authority problems that would result from UKM records bringing unauthorized headings into the database. And if we used another record besides the UKM, we’d have to delete our holdings from the UKM record, which is an annoyance and easy to forget. It seems clear to me that when a non-UKM record is available for a title, accepting the UKM record from PromptCat is counterproductive.
C. When only
UKM records are available, do those records tend to be any better than the ones
that co-exist with non-UKM records?
Yes.
Definitely. Unfortunately, it is hard to tell what one of these records looked like when PromptCat selected it—in many cases the records have been upgraded on OCLC since then. But if we could somehow accept UKM records only when there are no non-UKM records available in OCLC, it certainly appears that we would get some copy that would qualify as speed LC copy cat, some CC copy, and some pretty good member copy, as well as a still notable number of shoddy un-upgraded records. (This is certainly logical. It appears that the key may be whether the first OCLC member library to catalog the book enhances the UKM record or makes a new record. If the former, other libraries follow suit; if the latter, other libraries use that record and skip the UKM. This is surely an oversimplification.)
I asked Mary Ellen a question I had about the “freshness” of the UKM records PromptCat would be providing. I knew that they select an OCLC record for a particular book prior to and independent of our ordering it, via a process that involves the book’s vendor. I asked: do they then store that OCLC record in some sort of separate PromptCat database and not update it, so that eventually the new record is out of synch with the one in the WorldCat database? Or do they just keep track of the OCLC numbers of the records, and send the most updated versions of those OCLC records to us when we need them? It turns out it is the latter, fortunately—if the former were true, we’d be getting some awfully stale records. Thus, it should be possible to go through a VERY recent PromptCat report and check what the corresponding OCLC records look like. Then we’d get an idea of what we’d be getting for the “UKM-only” titles. We might find out that the records don’t look as good as they do two or three weeks later when verifiers pull them up upon re-searching.
In any case, though, it is important to note that many issues related to titles for which there are both UKM and non-UKM records disappear when it comes to “UKM-only” titles. It would be necessary much less often to overlay the record in cataloging. Deleting holdings would not normally be necessary. There is no question of whether our holdings symbol is on the most appropriate record.
Given, also, that letting PromptCat give us a UKM record means less re-searching work for the verifiers, I think we clearly come out ahead in going that direction when the UKM record is the only one available.
D. Close to
half the titles involved have both UKM and non-UKM records available, while the
others don’t. Thus, it appears that
either way we go, our decision will affect about half the titles positively and
half negatively. Let’s look at it in
terms of workflow.
If
we started receiving UKM titles, Order Services staff would have less work to
do, which is a good thing. They would
have to do less re-searching, since books that would previously have come in as
NC would now have records. Some of the
records would have been considerably upgraded since UKM created them, so there
would be a mix of MC, CC and LC copy (going by the criteria in Order Services’
"how to determine coding" document).
The MC and CC books would come to original catalogers or LAIV’s. Since the book would not have been re-searched in OS, it is likely that the cataloger would commonly re-search it. (This would happen whether there was a non-UKM record available or not, of course, because the cataloger wouldn’t know in advance.) Often, the cataloger would overlay the record on Acorn, either with a new non-UKM record or possibly with an upgraded version of the UKM record. If she didn’t use the UKM record PromptCat provided, she would have to de-update us on that record (or have someone else do it).
Some name headings arriving with UKM records would become a problem for the cataloger and the authorities unit, though not as often as they would have before the blind references problem was solved. Of course, in the case of “UKM-only” records, this can’t be helped.
Now let’s look at the current system. Order Services staff re-search books that would come with UKM records if we were accepting them. (If I understand Alice correctly, this usually requires only one actual search through Smartport, so the costs of doing this are not great.) They note on the record the date of re-searching and route the book according to the level of copy they find (or keep it and speed catalog it if possible). This means that they spend more time than they would have to if we accepted the UKM records. In regard to the UKM records for which there is no corresponding non-UKM record, they are thus wasting time doing work that could have been done by PromptCat—except that their re-searching provides catalogers with the most up-to-date version of the record and a date code verifying that the copy is in fact up-to-date. This is useful. (The verifiers could still re-search if the UKM records were received, but then we would run into multiple search charges—as we would down the line anyway if a cataloger did the re-searching.)
In regard to the UKM records for which there are corresponding non-UKM records—I think we’ve found that we will most often want to use the non-UKM records. This obviously makes re-searching a necessity, either by a verifier or a cataloger, and perhaps means we gain by not having PromptCat give us the UKM record in the first place.
One benefit of receiving UKM records is that at least a record with some content, rather than a nearly useless placeholder, is on Acorn from the beginning. Some UKM records are not very accurate—they may have the title wrong, in fact—but they’re generally preferable to the PromptCat pass-through records.
E. Cost
considerations
Here are some relevant cost figures from Solinet (I think these are as up-to-date as we have):
Z39.50 search/export (i.e. Smartport): $1.1948
PromptCat processing when we get a record: $2.0532
PromptCat processing when we don’t get a record: $.6380
Thus the price difference between accepting a UKM record and not accepting a UKM record is about $1.42.
If we do not accept the UKM record, re-searching the title
on OCLC via Smartport and exporting the record found will cost about $1.19 if
only one search is necessary, which I gather is usually the case. This would be less than that $1.42
figure. Furthermore, I think a
cataloger would be less likely to re-search OCLC herself down the road, if she
knew that the verifier had done re-searching quite recently, than she would if
we received a PromptCat record which was chosen two weeks or so longer
ago. So we could save some additional
searching fees—at the least, that might balance out the times when a verifier
needs to do two Smartport searches in re-searching.
PromptCat does update us on a record when they provide it to us, and that’s part of the benefit of the service, but it’s not a critical point, I think. If the record is not updated by a verifier via Smartport upon re-searching (it wouldn’t be except, I assume, when the record can be speed cataloged), then it can be updated via batch process for free down the line. Or, if a cataloger does look at the record on OCLC again, the cataloger can go ahead and update it at that point without additional cost.
I don’t think it makes a lot of difference to our OCLC fees whether we accept the UKM records or not. Of course, staff time is also a financial consideration, but matters related to that are discussed elsewhere in this report.
F. What I think
I think catalogers benefit from the current system, because theoretically the best record is provided to them, along with a date code that tells how recent it is and when it was found to be the best record. They do have to take care of updating, whereas they wouldn’t if PromptCat provided the records, but that seems a minor detail, especially considering that they won’t have to de-update UKM records which are not used locally.
Not receiving UKM records puts a burden on Order Services staff, but it does not seem to have been a particularly onerous one, and I think they may be in the perfect position to choose the record Vanderbilt will use for cataloging. PromptCat is biased in favor of scanty UKM records that we don’t find useful, and makes its choice of records quite a bit ahead of our cataloging the books. If re-searching is done by a cataloger after it gets to her—in effect doublechecking PromptCat’s work—then she may end up dealing with a book that should have gone to some other staff member. On the other hand, a verifier has her hands on the book immediately before a decision must be made about how to route it, and often not at all long before whoever gets it is going to catalog it. Thus the book should get to a cataloger at the right level, soon enough for her to be able to trust the record as it is on Acorn.
Not receiving UKM records does mean that only the barest excuse for a record sits in Acorn for a while before it is re-searched, but that doesn’t appear to have troubled anyone very much so far and in the absence of complaints, I think other concerns are more important.
The
Task Force composed a survey to send to a group of other academic libraries on
their cataloging organization and workflows.
Most of the libraries were chosen by Bryan from an ACRL report as
comparable to Vanderbilt in size.
Unfortunately,
we received responses from only 2 of the 14 libraries we sent the survey to
(the University of Cincinnati and Georgetown University). Even those responses are not especially
helpful. The respondent at Cincinnati
was unable to give us much help on many of the questions, and the one at
Georgetown skipped a couple of the questions entirely and responded pretty
minimally to the others. Frankly, we
kind of abandoned the survey in disappointment, and it didn’t contribute very
much to our deliberations, recommendations or report, but it is offered here
for some outside perspective on cataloging workflows. We have probably glossed over a useful nugget or two.
Below
are the original questions from the report, with the responses we
received. I’ve also incorporated
information we received earlier when we asked the head of cataloging at the
University of Tennessee-Knoxville a lot of questions about their
department. I’ve rearranged and split
up this information to make it fit under the question headings, though of
course she was not originally responding to the survey itself. We got much more from her than from the
other two libraries, and I believe her responses did spark a certain amount of
discussion in our task force meetings.
UT’s cataloging work is in general done by “lower” levels of staff than
ours is; professional catalogers apparently do much less actual cataloging than
ours do. We did not feel inclined to
recommend that Vanderbilt follow this lead aggressively, but their approach to
training paraprofessional staff probably merits some attention. If, as
envisioned in section 1 of the task force report, a shift in physical location
brings together our catalogers and acquisitions staff in one place, making
reorganization of cataloging duties a more fruitful prospect, we might find it
useful to learn more about UT’s experience with cross-training and how their
various cataloging “ranks” work together.
1. For what libraries/branches does your
department catalog? (By “department” we
mean a whole cataloging operation—for example, this might be a combination of
original cataloging, copy cataloging and serial cataloging units.)
UC: The music library catalogs its own items,
but is part of the library—the law school and medical school are separate
schools and have their own operations
GU: GU main campus libraries
UT: Not addressed. Presumably the department catalogs for most or all of the library
system.
2. Can you provide or estimate statistics for your cataloging production for the past year (or for another time period that is convenient)? If possible, can you divide these up between original cataloging, copy cataloging, and outsourced cataloging?
UC: I do not have the figures
GU: 32,000+/year (no outsourcing)
UT: 28,379 titles added in fiscal year
2001/2002. 12,131 came shelf ready from
Yankee and needed only occasional spot checks locally. Very little cataloging was done by
professional catalogers.
3. How many FTE cataloging staff are in your department? Are there any staff members outside your department who have cataloging responsibility (perhaps acquisitions staff members)?
UC: CDM has 4; Bib. Access has 5
GU:
9 FTE’s. 2 Area Studies Bibliographers in Collection Development Dept. spend
30% of their time performing cataloging.
UT: They have 25.5 paraprofessionals who do
cataloging, but 6 of them in Serials do very little actual cataloging. It looks as if the 25.5 includes
acquisitions staff as well as strictly cataloging staff; see no. 5 below for
information about the cataloging done by acquisitions staff. As for professionals: 2 professionals do most of the training,
review and cataloging. The Head of
Cataloging and the Head of Serials apparently do not catalog, though the
Serials Head trains catalogers. A
metadata librarian “works with non-MARC schema” and “spend[s] considerable time
developing cross walks between MARC and these schema,” but I’m guessing she
does little actual cataloging. A music
librarian catalogs music half-time.
4. What levels of cataloging personnel are defined in your department (e.g., professional cataloger, library associate, acquisitions staff member who does simple copy cataloging, student assistant)? How many of each category are in your department, and what percentage of their time is spent in cataloging?
UC: 2 with faculty rank, 2 more with MLS and not
faculty rank—rest are paraprofessionals, with many years of experience—students
are used for routine tasks
GU: 2 professional catalogers (60% of their time
doing cataloging
5
copy catalogers (60-70% of their time doing cataloging)
1
authority control specialist
Dept.
Head
UT: See above for some information related to
this. She also says, “The great
majority of the time of our librarian catalogers is spent doing review and
instruction of the non-librarian catalogers, their faculty duties, and their
duties as collection managers. They also get to deal with the complex
recataloging. At times, they are lucky
to produce two original records per week … Catalogers also plan and execute
training sessions …Staff assigned to cataloging unit DO do more cataloging
[than either librarian catalogers or acquisitions staff, I think she
means].” Apparently, while acquisitions
staff do cataloging as well, it makes up a considerably smaller portion of
their job than it does for non-professional cataloging staff—which makes sense.
5. How is your cataloging workload divided
among your staff? By subject or
material format? By level of cataloging
copy available? Please describe. Who does most of the original
cataloging? What level of copy (or
original) cataloging, if any, is done by staff members who also do acquisitions
work?
UC: Bib. Access does copy cat; CDM does upgrade,
original, problem records and database mgmt, and training of student help
GU: By both subjects and material formats
Catalogers
perform original cataloging, upgrades and enrichments.
Copy
catalogers perform copy cataloging.
UT: First off, the acquisitions staff members do “routine DLC and copy cataloging for shelf ready books … copy cataloging for firm order materials, though that is more routinely routed to the cataloging staff … [and] are asked to catalog 2 books each month from the backlog, though that has not been strictly enforced. They do original cataloging (from a workform