The OPAC as Portal - was [WEB4LIB] Amazon.com's purchase circles and the digital library

Tony Barry me at Tony-Barry.emu.id.au
Tue Oct 19 20:59:35 EDT 1999 

At 02:02 PM -0700 1/10/99, Prentiss Riddle wrote:
>when will the digital library go beyond the "card" catalog
>model and start implementing friendly features like this?

Nobody has taken up on Prentice's comment above.

Is the OPAC a portal?

We need to decide what the OPAC is for. Is it and add on to the
inventory/circulation system or a tool to find information?

Is it a tool to search for what the library physically holds or pays
to access or its it a tool to search for other kinds of documentary
information?

Should it be a tool only for searching for documents or information
more generally?

Why do we have library web servers and the OPAC and think of them often
as two separate things?

Should we create gateways into the OPAC from the library web server
and regard the web server as the prime source of information where the
OPAC is a legacy application or should the OPAc become the search
engine for the library web site?

Lets look at some possibilities. One key to an expanded role for the
OPAC is authentication. The OPAC can custamise the service for
individual user if it knows who it is. Libraries have missed this step
as they still think of the OPAC as a general public service in the
library like a card catalogue. Another key is dynamic information
which changes according to circumstances such as time of day and the
users past behavior. I'm led to this view by looking at what the
portals and doing and what many of the bookshops like Amazon are doing.

This is what I'd like to see (I'll indicate hypertext links by '_') -

GREETING
--------
When I log into the OPAC I want it to greet me with immediate
information which may be useful to me if I am to visit the library
such as -

	Hullo Tony. Today is Wednesday 20th October and the time is 10am.

	Library branch A is open and will close at 7pm
	Library branch B is closed and will open at 1pm closing at 5pm
	Library branch C is closed until 9am Monday.

	Rain is forecast.

	You have 8 books on loan.
	Dirac,P Quantum Mechanics is due back on Friday. Do you want to
	  _renew_ it?

	There are 15 _new books_ which might interest you.

	Some _journal issues_ you asked about have arrived.

SEARCH RESTRICTIONS
-------------------
That then leads me to the main page of the OPAC/Portal?Web site for
the library. If I choose to search the OPAC if allows me to restrict
my search not just by location or form but by -

	What I have borrowed in the past.

	Subject headings and classifications I have specifies in
	this and previous sessions.

	Whether they have borrowed by members of a given faculty or
	department.

	Rated above a certain level (see below)

	Date of acquisition


REVIEWING AND RATING OF BOOKS
-----------------------------
If I borrow a book and return it I want the system to prompt me next
time I log in and ask me if I would like to write a review and also to
provide a rating for the book. In searching the reviews would be
available and average and range of the ratings.


SEARCH FACILITIES
-----------------
When I see an item I like rather than just hypertext links to other
books which share the same headings I want a "More like this" facility
which accumulates as I mark the items which I like and drop those I
don't.

The matching algorithm would be based not just on shared headings but
also on a dynamic analysis of borrowing patterns. If two books are
borrowed by the same person there is an increased likelihood that
they are related. Another possibility is an equivalent of the Amazon
facility eg "People who borrowed this book also borrowed "XYZ".

In the process of doing this I would want to be able to mark subjects
headings, authors and classifications as being of interest to me so
than in the future I could use them to restrict or focus my searching
as described above.

MATERIAL NOT HELD
-----------------
The OPAC should not just be limited to material held physically or for
which there is paid online access. For instance failed searches for
monographs could give the option of searching an online bookshop. For
that matter any book could have a hypertext link to an online bookshop
in case the user wanted to buy the book. The Amazon Associates program
would then generate revenue for the library from the sale.

Entries for serial titles NOT held could be added to the OPAC to link
to document delivery or ILL services.

But why limit it to this? Many approval plan/blanket order vendors
will supply MARC records for material judged to match the library's
needs. Load these with a hypertext link to a form which lets the end
user shift to the bookshop to buy it themself or provide a rating for
the book from a range of -

	I need to read this book immediately

	I will borrow this book if it is acquired

	I may in the future borrow this book

	Somebody will like this book
		[provide link so they can email a colleague]

	This may be of interest to somebody in the future

	This book is of no interest to the library

On the basis of the responses, collected automatically the library
then decides to buy or borrow. A mechaanism like this should greatly simplify
decisions on book purchases as there will be a clearer indication of demand.

NEW BOOKS
---------
I mentioned at the start a notification of new books arrival. This
could be based on -

	Subjects or classifications I had specified

	"I'll have what he's having" eg books of interest to people
	who have a similar borrowing pattern to me

	Books I have said I wish or may wish to borrow.

INTERNET CONTENT
----------------
The internet is too dynamic for a single organization to catalogue but
a tiny part of those things that may be of interest to their users.
There are possibilities should somebody provide such a service, to get
a feed of MARC records for internet content based on blanket order
criteria but why not involve the users? Catalogers take documents and
assign multiple subject headings to them. You could invert this. A
dummy bibliographic heading for each subject heading could act like
the internet equivalent of a generic pamphlet box where the users
could attach URLs. Then the users would be attaching document to the
subject headings. The librarians would control the subject vocabulary
and the users the selection of documents in which they are interested.
Both would be doing what they know best.

In addition I would want dynamic links from the subjects to
searchengines to search on that topic.

ALLIED INFORMATION
------------------
The internet has other things besides bookshops. Other links would
be possible from an OPAC.  Link from -

	Authors to their home page

	Authors to sites discussing their work

	Publishers to the sites

	Authors to their email addresses

	Catalogue entries to book reviews of the book

	Call number to a floor plan showing roughly where the book
	is located

PIE IN THE SKY?
---------------
ILMS vendors are lucky to sell in the hundreds rather than the tens of
millions than major shrink-wrapped vendors sell. It is unlikely that
they can turn round and do this. Some options are -

	Treat the OPAC as a legacy application, do some data mining
	and use other software.

	Libraries, using the open source model could do it them
	selves See the Avanti project
	<http://nsls1.nslsilus.org/~schlumpf/avanti/project-outline.html>
	and gnujake <http://gnujake.med.yale.edu/about.html>

Once we stop saying the OPAC is about collections we have to
ask what IS it for but can we make a silk purse out of a
sow's ear?

Tony

- - - - - - - - - - - - - - - -
phone  +61 2 6241 7659
mailto:me at Tony-Barry.emu.id.au
http://purl.oclc.org/NET/Tony.Barry

More information about the Web4lib mailing list