Interview: Tracey Robinson
Tracey Robinson is the head of the Harvard University Library Office for Information Systems (OIS). As such, she is the chief executive for the implementation of Aleph, the new integrated library system that will replace the HOLLIS catalog in July 2002. Robinson came to the Harvard Library system as a systems librarian in 1980. She was, in her own words, "intimately involved" with the development of the original HOLLIS system.
On November 15, 2000, Harvard entered into an agreement with the Ex Libris company (see Library Notes at
http://hul.harvard.edu/publications/library_notes/1299/aleph.html) for the purchase and installation of its Aleph system. Tracey Robinson was interviewed for Library Notes on December 14 specifically to provide an inside view and a general update for Harvard librarians and library staff on the Aleph implementation.
Everyone has the same question: one year into the contract, how is the
Aleph implementation going?
Well, we've got six months before we "go live" on Aleph. We've accomplished
a great deal but there is still much to do. There are more than a hundred
Harvard librarians involved in configuring Aleph and preparing Harvard for
using the new system. Broadly speaking, we're working very hard in six
Conversion. We have to migrate all active data from HOLLIS into Aleph. This
is clearly a critical task.
Configuring the system. The system must be setup to reflect Harvard
policies, practices, and organizational structures.
Desktop and server operations. We've been learning how to configure both
the desktop client software and the server software. We're also developing
procedures that will allow us to distribute, install, and maintain the
desktop clients, which will be used by library staff: a total of about
1,200 computers located throughout the Harvard libraries.
Training. It's a massive effort to get all Harvard librarians trained.
Curriculum is being developed now and classes will begin in the spring.
Training for students, faculty, and other "end users" will come later and
will be handled locally.
System developments by Ex Libris. We've contracted with Ex Libris for a
long list of enhancements. We saw the first set of these this past summer.
The second round of developments are being installed this month and a third
set of changes will be done by February or early March. We're all anxious
to get our hands on the next release of the system to see how well they
have implemented our specifications.
System developments in OIS. There are some developments we'll need to
program ourselves. The biggest local development, of course, is the design
of the new online catalog interface. The new web-based OPAC will be a
dramatic improvement over the old HOLLIS catalog in terms of design, ease
of use, and functionality. Demos of the new OPAC are planned for later this
spring. Other local developments include setting up a data feed between
Aleph and HD for processing recall requests. There are also electronic
"links" to other Harvard systems that must be established such as accounts
payable, accounts receivable, student term bill, the university PIN
service, and the central patron database.
Will you be on time?
Yes, we fully expect to go live on time. There will be some features that
won't be ready for Day One but given the energy that people are putting
into it and the oversight provided by the steering committee, I am
confident that the system will come up in July.
When is Day One?
July 1, 2002. The plan is to bring HOLLIS down-with the exception of
circulation-in the middle of June. Circulation will continue until right up
to the last weekend in June. And then, on Friday, June 28, we'll shut down
HOLLIS for good and bring up Aleph on Monday morning.
Will end users need to know the name "Aleph"?
Aleph is the vendor's name for the system. We won't use Aleph as a public
system name. There have been recent discussions about what to name the new
system and we expect a decision from University Library Council soon.
What are the gains and losses of going from a home-grown system like our
current HOLLIS catalog to a system like Aleph that is, in a way, bought off
The perceived losses are mostly in terms of independence and control. We
need to negotiate with the vendor, and collaborate with other customers, to
prioritize new developments. But the gains are significant. If we had
chosen to build our own system again, the cost of doing it and the time
that it would have taken would have been enormous. It would have taken
many, many, many years and the project would have distracted us from
developments in digital library services.
Transitioning from HOLLIS to a new system will undoubtedly feel like a loss
to many people but we should all keep in mind that Aleph will fit much
better into the current technical environment than HOLLIS ever could. It
will provide features that offer long-awaited improvements such as a new
web-based online catalog. Among other things, these will allow patrons to
setup profiles and access personal library account information, online MARC
documentation for staff, facilities for importing bibliographic records in
real-time, predictive serial check-in, the ability to email orders and
claims, just to name a few. Finally, there will be future gains because
the new technical environment will allow for a more rapid rate of
development than what we were used to in the HOLLIS mainframe environment.
When the agreement was made with Ex Libris, six specific areas of
enhancements were to be addressed, by contract, for the Harvard
implementation. How are those going?
We will see most of the enhancements in the next two months. They include:
Improvements to indexing. Our database is one of the largest ever converted
to Aleph. Indexing is the single most important feature of a library
databaseit must work well or patrons won't be able to do their research
and staff won't be able to do their work. We spent a lot of time on
specifications for index improvements to ensure that searching our large
database will work well and make sense.
Improvement to the input and display for MARC holdings data. This is one of
the really hard ones. We're not going to see holdings until February, and
there's a lot of frustration right now on the serials team since they've
not been able to see the improvements related to serial holdings. Our
holdings are vast and displaying holdings coherently for patrons is a major
challenge. Data in holdings records may also be used for serial processing
and check-inproviding the holdings records have been encoded correctly.
Is that a particularly intractable place?
At the moment there is serious concern because Aleph supports automated
predictive check-in and
we are not quite ready for it. Predictive check-in allows the system to
say, "Okay, I know something about the publication pattern and timing of
this serial, so I can predict what you expect to receive and when you
expect to receive it. I'll set up check-in screens for you with your
expected issues." But predictive check-in relies on serial patterns, which
live in the holdings records. Although some libraries at Harvard have been
building MARC holdings records for the last fifteen years, not all
departments have done it comprehensively or have kept the records up to
date. And because that pattern data hasn't been used by a computer, it has
not been tested. So, there's an unexpected amount of conversion work for
holdings before predictive check-in will work.
Back to our list: enhancements to make Aleph compatible with Harvard's
33-digit chart of accounts.
The chart of accounts and several other things in the accounting area are
due in January, so we will see them soon. Those changes are straightforward
and I'm confident that we'll be able to work well with what they've done
Improvements to staff/client interfaces (for the circulation, acquisitions,
and cataloging modules).
This is an area where we're going to see some delay-and we've already put
out public announcements about this. The Aleph staff/client interfacesthe
"look and feel"-need improvement. They are difficult to learn and seem
inefficient because of the excessive number of windows and inconsistencies
between the different functional modules. Last year, Ex Libris (wisely)
hired a designer who has been working on overhauling the staff client
interfaces. He has developed some
very interesting prototypes. Although the client interface for circulation
will be done before we go live, the client interfaces for all the technical
services modules (such as cataloging, acquisitions, and serials) will be
delayed until next year.
According to the contract, there are meant to be new features relative to
These are not really new features for usbut rather adding to Aleph,
features we have in HOLLIS that allow patrons to request items from HD.
There will be some improvements over HOLLIS because we will add a process
to feed information from HD back to Aleph, when materials are returned to
the Depository, thus improving our ability to track and locate items housed
How about support for non-Roman bibliographic data?
Non-Roman is a very important development for Harvard. In January we'll see the developments related to the support for Chinese, Japanese, and Korean characters. Aleph will support other non-Roman scripts as well (Hebrew, Arabic, Cyrillic, Greek, etc.). The ability to support the non-Roman characters will provide a new opportunity for catalogers creating records in these languages. We will be loading something in the range of 300,000 records that have vernacular data in Chinese, Japanese, and Korean. Catalogers will be able to create and maintain records including vernacular characters in Alepha major advance. These records will be loaded as soon as Aleph comes up and will be available in the catalog before the fall semester.
The next two months include quite a number of critical moments.
The first big milestone is loading all 9 million bibliographic records-and
all of the orders, items, and holdings records associated with them. And
then there's the process of indexing the full database, which we expect to
be very time-consuming. By the end of January, we'll have a better sense
of how long it takes to load and index the entire Harvard file. Once the
data is loaded and indexed, we will perform a series of performance or
"scale" tests. This step is critical one for us given the size of our
database. Ex Libris does not have much experience with running a database
the size of Harvard's.
Not many vendors would.
We have plans for doing at least two scale tests: one in February and
another one in April. From these
tests, we will learn how to tune or re-configure the system and whether we
have sufficient hardware to support good performance and reasonable
Another big milestone is the beginning of the training program. Training
will start in early April. The most intensive training period will be April
through July, but we expect an elevated level of training classes to
continue well into the fall.
Who is actually handling the training?
Julie Wetherill and Cathy Conroy are coordinating the training program
along with the Training Advisory Group. They are working with forty
volunteer trainers from the libraries who are developing the training
curriculum and who will be carrying out the training starting in April.
That's terrific. You've had a great response from the entire library community.
Yes, the level of involvement by library staff has been really impressive.
I've heard and seen a wide range of responses from library staff in the
past year. Of course, there is a lot of anxiety, but also curiosity about
what it all means and where we're going. There's anticipation, a certain
amount of dread, and certainly a bit of resistance. But there's also
excitement. The opportunity to rethink the way we do our workand to
reinvent the way things are doneis surprisingly rare. I think that,
particularly among the implementation teams, there's a great deal of
satisfaction that goes along with the frustration and the hard work.
Working with Harvard librarians is such a pleasure because they approach
this kind of challenge with a great deal of intelligence as well as common
sense. They're very creative but they're also very practical in how they
approach things. And they're very eager but they're also patient. All of
these skills will contribute to the success of the Aleph implementation.
Table of Contents |