oss4lib About - Contact    History 
Listserv         Projects
Readings         Submit  


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: open source and librarianship



Eric Lease Morgan <eric_morgan@ncsu.edu> writes:
> Thank you for asking. In a sentence, MyLibrary, specifically,
> MyLibrary@NCState is a user-driven, customizable interface to a
> library's collection of information resources. It operates very much
> like a customizable portal, but includes features that the typical
> portals do not include such as human interaction. I have written a
> number of documents about MyLibrary@NCState. See:
> 
>   http://my.lib.ncsu.edu/about/
> 
> You can try the system as a guest by following the instructions here:
> 
>   http:/my.lib.ncsu.edu/

Thanks!  I took a look at the site (and read your paper); it seems to
be not only an interface to a libraries collection of resources, but a
kind of generalized bookmarks system in which some of the links are
chosen for the user based on her specialization.  And the easy path to
human interaction is a wonderful idea!  But yes, you're right (below)
when you say that MyLibrary is solving a different problem than the
one I had in mind.

There are two reasons I focus on electronic card catalogs: first,
books and other printed matter are what a library actually *has* --
that is, they are they raison d'etre of the library's physical
existence.  A library may also have electronic resources, in some
loose sense of the word "have", but if the library is doing its job
right, those resources are available to anyone on the Internet, with
no limit on the number of copies viewable at one time.  Thus, though a
library may have some custodial duties pertaining to the maintenance
of electronic documents (including maintenance of the search engines
that lead people to those documents), when it comes to online content
a library is still essentially a transient middleman, needing no
relationship and only very minor interactions with the user.  (And
indeed, most of the online documents available from the MyLibrary
guest account page were actually offsite -- NCSU has links to them,
but does not maintain them).

And reason number two: most -- no, make that all -- of the online card
catalog systems I've seen just plain suck.  I mean, *really badly*, in
ways that one wouldn't have thought possible.  I look at them and just
*know* we can do better, that (specifically) a Free Software solution
could do much, much better than these systems, saving libraries both
money and headache.

For example, my two housemates use the University of Chicago library
system every day.  One is a graduate student, the other is actually an
employee of the university's Regenstein library, involved in
categorizing and filing.  I had Karen (the grad student) give me a
tour of their Web interface to the card catalog, which she says is the
same as what you get in the library.

In this state-of-the-art system, which apparently the university paid
Ameritech real money for, it is _not possible_ to narrow a search on
more than one field at a time.  That is, you can search on author, or
title, but not both.  Dumbfounded, I searched the online help with her
until we found a description of a search syntax that appeared to give
one a (limited) ability to search multiple fields at once.  Understand
that this was a complex, obscure syntax, buried in documentation that
no user would ever read anyway, and that Karen -- a highly
intelligent, computer-savvy person -- had to try it several different
ways before she could even get the system to accept the search.  At
which point, even this limited multi-field method turned out to be
broken so severely that it's essentially unusable.

Even with the approved single-field search method, the results come
back in a way that's less than useful.  Specifically, the result list
usually doesn't give enough information about each match for Karen to
tell which ones she's interested in.  Instead, she has to click on
each one and read the details.  And it's all the more offensive
because most WWW search engines get this one right nowadays -- do a
search at altavista, google, wherever, and the result list shows not
only basic item identification information (title, author, etc), but
also a tiny blurb summarizing what the item is (often this blurb is
summarized automatically from much lengthier detailed item information
in the system).  Karen -- and, I suspect, most library users -- would
save vast amounts of time if she had that sort of information right
there on the results list.

I could go on, but I think you've all heard the tune before.  After
ten minutes using this system I already had in mind a list of obvious
(and easy to implement) improvements longer than I could remember.
Not that I had the source code, of course. :-(

As for my other housemate, David, the Regenstein employee, he says the
system they use at work crashes approximately ten times a day, and has
a lousy interface (I haven't gone in to see the interface myself yet,
so this is second-hand, though I believe him completely of course).

These sorts of problems are not unique to the University of Chicago.
May I hit closer to home, honesty being more useful than gentleness
right now?  I am not a regular user of the NCSU library system, but I
followed the link from MyLibrary to the NCSU online card catalog, and
saw that it also allows only single-field searching.  Bleargh.

So I did an author search (I admit it, I searched on my own name
"Fogel" :-) ).  When the results came up, they were too many to fit on
one page, so it listed "Fogel, Aaron" through "Fogel, David".  But
nowhere on that first page did it indicate how many *total* results
had been returned, nor how many following pages of results there were!
The basic user desire to know how big a task lies ahead could so
easily have been satisfied, but wasn't.

When I see a niche so grossly underfilled, and so open to the
possibility of a collaborative effort to fill it, I get all excited
and weepy :-).  But has this problem already been solved?  I don't
know.  None of the libraries I use are actually running an Open Source
search system, but then again I only use a few libraries.  Is there
already category-killer software out there and it's a mere matter of
spreading the word?  Or if not, what kind of interest is there on this
list in creating such software?

> If I understand you correctly, then I believe MyLibrary is not
> necessarily what you are invisioning. What you might be thinking about
> is a common interface to searching library collections such as card
> catalogs. There is such a protocol, an ANSI standard, in place, Z39.50.
> There are also a number of open source clients and a few servers, but
> most libraries are at the mercy of "vendors" who sell and support "card
> catalog" systems. I believe I can speak for most people here when I say
> that we feel the support from these vendors is much like the support one
> gets from a monopoly. 

That's exactly what I've heard from other librarians, yah. :-(

What is your opinion of Z39.50, and of these OSS clients and servers?
Obviously, I'm on very unfamiliar territory.  Are any of them ready
for prime time?

I actually don't think it's such a good idea to concentrate on
designing an interface with the intention of its being common among
all libraries someday.  If a system is designed from the start to
please everybody, there's a danger of its settling on a lowest
acceptable denominator.  Rather, make a *good* interface that pleases
one or two libraries initially, and if it's really so good others will
naturally adopt it over time, making their own suggestions and
improvements.  Interfaces shouldn't be specified top-down, but instead
should evolve, because they are so sensitive to real-world usage
patterns.  Of course an initial design is necessary, and as long as
there's an initial design some thought ought to be given to it, but
mainly so the evolutionary process has something to work from.

> You can see a list of some of these Z39.50
> implementations here:
> 
>   http://info.med.yale.edu/library/oss4lib/projects/welcome.php3

Thanks!  I looked it over, didn't immediately recognize any names, but
then again this is a new area for me.  What software on that list is
good and non-vaporous? :-)

Best,
-Karl


SourceForge Logo © Copyright 1999-2005, The oss4lib Community, except for readings and comments, which are owned by their posters.
oss4lib is graciously hosted by the good folks at sourceforge.net.
Site URL: http://oss4lib.org/ Questions or comments to maintainers.


library