Overabundance of riches (database lists)

[Stacy Pober]

I'm musing over a problem that is part design, part technical and partly
philosophical in nature.  Just jump in on whichever points appeal:

We already have a page listing the most-used (or what we consider the
most-useful) databases that our library provides.  This lists most of
the aggregators front-ends, but does not provide deep links in most
cases - so you can see that we have Dialog at CARL, for example, but not
that Dialog at CARL includes SPORTdiscus, a very useful physical education
and sports database.  We do "deep link" to a few selected databases, but
I suspect that most users don't really look into the large selection
that is in some of our aggregate collections and thus may be missing
some resources that might be quite useful to them. 

I want to try to create an alphabetical list of most of our subscription
databases, including the individual databases inside the aggregate
sites,  And I'm having a little trouble deciding how to organize this on
the web and how much information to provide.  

Looking on the web at other libraries,  I see several common approaches
to this situation: 

- a table (or "invisible table" of columns) that provides links to each
database along with some very brief information.  The categories of info
included in these lists  often include the coverage dates, type of data
[whether it's a full-text or an index database), sometimes the vendor
name, and often a link to more information. 

- a dictionary list with a wordier text description that details the
scope, subject areas, and so forth.

- a subject-tree or alphabetical list with almost NO information - just
the names of each database and in some cases icons to indicate whether
each resource is limited to on-campus users or free on the web. 
Sometimes other icons denote whether the database is full-text or
bibliographic in nature.

I can see pros and cons in all of these.  I started working on a list of
links to the databases that are included in the Dialog at CARL basic
collection.  I first excised the databases that are actually electronic
versions of individual publications, such as the New York Times, as we
already maintain a seperate web resource listing all the full-text
periodicals we get through vendors (yep, all 12,000+ links - criminey,
that is a PITA to maintain... )

My first thought was to include four columns:  
Database Name (and hyperlink to the database)
Data Type (short notation as to whether it's full-text, index, or
whatever)
Date Range
Link to a longer description on the vendor's website

Here's where I'm running into trouble.  Dialog at CARL has hundreds of
databases in it, but I don't seem to be able to link to their internal
database descriptions.  The links to the actual databases - which DO
work - are in the form: 
http://dialog.carl.org:3005/cgi-bin/cw_cgi?7005+REDIRX+setDatabase_6292

but the link to the description of that database is: 
http://dialog.carl.org:3005/cgi-bin/cw_cgi?describe+25577+6292

and when I invoke that one from outside of the database, it just hangs. 
I am not a cgi ace, but I'm assuming that the problem may be that the
database URL includes some provision for recognition of our IP address,
whereas the database description link does not.  

The vendor provided us with the internal links to the individual
databases, but the description links are something I just copied from
their pages.  I tried fooling around with the description links to see
if I could somehow insert the magic REDIRX coding and make it work, but
I haven't quite figured that out.  And right now, I don't have time to
write - or even copy from the site - all the database descriptions to
maintain them locally.  So I'm thinking about other approaches.  

Incidentally, I've been noticing how variable providers are in giving
customers and other users information on the databases they are
providing.  One of the reasons I'd like to link to the internal
descriptions on the Dialog at CARL site is that they are outstanding.  They
cover date, data type, and provide a good overview of each database.  HW
Wilson also does a good job of providing informative databases
descriptions.  On the other end of the spectrum, Lexis-Nexis Academic
Universe, another popular aggregator, has a source list that has
literally hundreds of dead-end links for resources that either don't
exist or cannot be successfully searched using their own links.

Anyway, this is leading me into the mire of:  How much information is
too much?  I know from my own surfing that there are times when
terseness and simplicity in presentation is much appreciated.  But some
of the databases have names which are fairly cryptic as to their
content, so I feel they may never be underutilized without some
descriptive information provided.  

For some databses, name, date range and data type is plenty:  Africa
News, for example, is pretty well self-explanatory.  But then there's
AGRIS International or Eventline or Kompass Canada which are far more
ambiguously titled.  

Another thing I've been thinking about is multiple access points for the
same database.  In Web Site Usability: A Designer's Guide, one point
they made about utility was that having several links to the same
resource, each with different wording, increased the possibility that
people looking for the information in there would find it.  That was
something I'd never thought of before.  So now I'm considering things
like listing some of the more ambiguously named databases under both
their official names and under a subject-oriented title, such as using
"Physics Database" as a second link for the SPIN database.   

What are you folks doing to enhance access to these resources?  If
you're happy with what you've got, or conversely, made mistakes that I
can avoid, please tell me!  

-- 
Stacy Pober, Information Alchemist
Manhattan College Libraries
spober at manhattan.edu
http://www.manhattan.edu/library/
List of our fulltext journals (non-IP-filtered version):
http://web1.manhattan.edu/library/journals/jinfo.html