YOUR RESPONSES to Which database to search

Thu May 12 17:25:05 EDT 2005

Colleagues

As Robert's and my e-mail indicates, I think that there is a desire to move 
away from subject listings of databases.  Unfortunately, in the absence of 
true federated searching of all databases, various problems arise, such as 
indexing.  The following are the responses to the original query I sent out 
last Thursday, I believe.  I particularly liked Bret Parker's response, 
tying my inquiry into the "stupid users" (not so stupid!) and web presence 
threads on the web4lib listserv at the time.  So while there appears to be 
a desire to do something like this, there are difficulties which must be 
overcome.

See below for responses:

*****************************************************************************************

I don't think that searching a list of databases is going to be very
effective.  I'm guessing that instead of searching for "business" they
will end up searching for "customer relationship management" or whatever
their specific topic is.

I just divided my list of databases into major subject categories (keep
the list of subject short, so a user can easily browse it).  My list is
generated by a PHP/MySQL application, but you could just do it with
plain html pages:

http://library.northwestu.edu/resources/subjects/

Adam

**********************************************************************************

1. Do any of your libraries do this?  If so, can you give me a URL pointing 
to the pertinent URL that I can look at?
You might want to look at our E-Resource Locator
http://www.library.upenn.edu/cgi-bin/res/sr.cgi

and this beta federated search tool

http://www.library.upenn.edu/federated/

2. What software does you library use to accomplish this purpose, or is it 
home-grown java?
Homegrown
3.  Is your OPAC configured to provide this service?  If yes, is there an 
accessible URL?
No
4. What keywords describing the databases are used and indexed, and how do 
you arrive at the pertinent words and import them into the database?
Initial coding by community of interest and type from same list that 
displays in on the search form. Non-displaying keywords can be added by 
librarians into a web based template, no controlled vocabulary.

Regards,

Danianne

-- 
Danianne Mizzy
Asst. Head, Engineering Library University of Pennsylvania
217 Towne Building
danianne at seas.upenn.edu
(215) 898-5700

*****************************************************************************************

This isn't exactly what you're asking for (and maybe
you already have something like this) but another
option is creating a subject index available directly
from the library website.

It's not a database in the traditional sense, but a
subject listing with links directly to the databases.
Unless I am experience end-of-week burnout, this seems
to accomplish what you've described in your email. In
my experience at these two libraries, patrons (and
librarians) have found this to be a handy tool.

http://brooklynpubliclibrary.org/eresources.jsp

http://qcpages.qc.cuny.edu/Library/online/dbasesubject.html

Hope that is of some use!

Have a great weekend.

Kara Masciangelo
www.infomascia.com

************************************************************************************

I gave the database approach a lot of thought when we wanted to make our 
electronic databases more visible to customers.

After examining some articles written at the time and the number of 
databases we had (less than 30), it did not seem right for us. Given recent 
discussion on WEB4LIB, which you might want to review (under the subject 
line "Library web presence" or "In defense of stupid..."), I am wondering 
how much your end user's experience would be enhanced or not by 'yet 
another' database to find library resources.

Not the Google approach, but aiming at user friendliness, we created a 
directory type structure. Since our 1-ton gorilla database was InfoTrac 
OneFile (at the time, we had "General Reference Center Gold" -- a 
practically meaningless title to patrons and without any branding power), 
we wanted users to discover that a "general" tool would help them on  a 
subject-specific topic.

So, rather than drop that database into the easily overlooked category of 
"General" or "Reference," we placed that and other similar 
multidisciplinary databases (e.g., Biography Resource Center, NewsBank, 
etc.) all over the pages leading people to databases, under labels that 
would trigger "want that" responses closer to the scent of the search at hand.

By comparing directory driven portals of the day (circa 2002, I think), I 
came up with about 31 some-odd broad topic headings that all the databases 
would hang from, plus some media-type or language-specific headings 
(magazines online, newspapers online, and espanol). We sprinkled web 
resources, with the databases, so the user experience was one-stop 
shopping. (I say, "we" but this was largely a solo production in its early 
stages. So, what you don't like, please don't blame on the committee!)

The end result is, of course, a lot of coding because the general-purpose 
databases will be repeated on many pages. The method of implementation, 
however, was XML, XSLT, done on a publishing workstation built largely of 
open-source tools (text editor and InstantSaxon at the core.)  Then the 
static pages are uploaded to the server.

I am still wedded to the described concept, though the implementation is 
morphing for various reasons. Its complete reincarnation is not yet ready 
for showtime, but what has been created, followed similar methods (and 
involved the work of many, unlike the first production).

The biggest downside is the need to "republish" *all* the pages that are 
touched by title changes/description changes when one wants to update or 
delete a cancelled subscription. The process is greatly sped up by the XML, 
XSLT manner of publishing. That said, there is a learning curve for 
learning to use this Swiss-army knife of web publishing. URL changes are 
easier to maintain as all URLs in the html are persistent. All links to a 
database point to a redirection script. If and when the URL changes, only 
the script file needs to be updated (and the proxy server config, if the 
resource is available to remote users). The scores of pointers to the 
resource go unchanged, in the event of a URL change.

At our web site http://www.stockton.lib.ca.us , to get an idea of what I am 
talking about, follow the link to "Find:  ... Databases" (just under the 
photos of the Kids), and choose the option in the upper right navigation 
bar "Databases, by subject." And that shows you the old screen with the 35 
plus categories.

Another approach that seems useful and attracts the user by "topic" and 
thereby appeals to the **browser** in every customer is implemented further 
down the main page of our web site (right side) under  "Hot Topics." Follow 
just about any topic, and we tell the customer about InfoTrac OneFile, 
Ancestry, NewsBank (or whatever database is most pertinent to the topic) 
AND the Library Online Catalog--with canned subject searches into our 
iBistro pac, encoded into the page, under less dyspeptic headings, than 
LCSH offers. See the Genealogy pages  for example, 
http://www.stockton.lib.ca.us/subjects/geneal/ , and you see a link in the 
right navigation bar, "Genealogy--Getting Started" that hides an iBistro 
subject keyword search which attempts the equivalent of this LCSH subject 
search: "United States--Genealogy--Handbooks, manuals, etc."  (granted, 
with Unicorn it is not quite precise, but good enough for the shotgun 
approach).

Don't know if this is useful for the audience in your setting or for the 
resources you have, but it is a similar activity, public or academic. Our 
project was possible largely because my supervisor gave me the time to 
develop it and it was reasonable to do with the few databases we had. For 
the in development pages, that have a richer selection of web links, a team 
of librarians coded the XML records describing each web link (The genealogy 
example.) Given their other responsibilities, that was a considerable 
resource commitment, as will be the maintenance of all those links they picked.

Bret Parker
Stockton-San Joaquin County Public Library
Stockton, CA
   http://www.stockton.lib.ca.us
-- 

*********************************************************************************

NC State is working a concept which is really intriguing. It's planned to 
go live this fall, but Tito Sierra presented a preview at a recent conference.

In Search of the Single Search Box: Building a "First-step" Library Search Tool

Presentation:
http://www.diglib.org/forums/spring2005/presentations/morris0504.htm

Video Demo:
http://www.lib.ncsu.edu/dli/staff/tsierra/dlf-ncsu-qs-demo.wmv

Abstract:
Libraries are under increasing pressure to provide users with a single 
search box that provides access to the diverse set of content and services 
available through the library website. Neither library catalogs nor generic 
Web site search tools meet this need directly. Metasearch, while promising, 
is still generally characterized by slowness, incompleteness in coverage, 
and confusing result sets. At NC State, an analysis of library Web site 
search logs indicated that a large percentage of user-submitted search 
terms target similar classes of content (e.g., database names, journal 
titles, library information) to which the library could readily provide a 
direct link. Concurrent with implementing a next generation metasearch 
tool, NCSU Libraries is developing a new Web site search tool designed to 
provide users with quick and comfortable access to distributed silos of 
library content. A "sponsored-links" component of this tool connects users 
to relevant high-use library resources and information. A subject- 
identification component provides contextual links to subject resource 
guides. Integrated results from ancillary local indexes enable use of the 
tool as a "first-step" in library search. This presentation will describe 
an in-house solution based on open source tools such as Nutch and SWISH-E. 
Challenges of current development will be discussed and future development 
directions will be outlined.

We have what we call our Research Advisor:

http://www.library.georgetown.edu/advisor/

There are drop-downs on our home page that allow you to select a category 
or topic before you get to this page, as well (under the "Databases" section):

http://www.library.georgetown.edu

As you can see by the URLs on the subject pages, it's done with 
ColdFusion.  This was in place before I got here, and I think the 
University IT folks did it for us, and it includes an administrative module 
to edit information about the databases (which is handy when you learn from 
a patron that a vendor has changed URLs and not bothered to notify you!).

There's a search function on the lower left of the Advisor page (the first 
link above).  I've never actually used it, and I don't know how much use it 
does get, but I don't think--based on what little I know about 
ColdFusion--that's it's *too* much extra work to add that search 
function.  You've done the hard part of setting up the database...

Hope this helps,

Nina

Nina McHale
Reference Librarian/Web Services Coordinator
Lauinger Library
Georgetown University

******************************************************************************

Mark:

TTU has such an interface, http://library.ttu.edu/ul/databases/, which has
been used for about 3 years now.

This is a Perl script that I wrote, and the the "database of databases" is
in MySQL. The database (in HTML table format) is attached.

This was meant to be a temporary stopgap solution to replace an
unmaintainable static list, pending our implementation of Metalib. Metalib
has been significantly delayed, but supposedly it's right around the corner,
so I haven't made much effort to make improvements to this.

~~~~~~~~~~~~~~~~~~~~~~~~~~~
Randy Norwood
Web  Manager
Texas Tech University Libraries
Office: 806-742-2238  x236
Fax: 806-742-8669
E-mail: randy.norwood at ttu.edu

*****************************************************************************

At Ithaca College, we currently are using the following:

http://www.ithacalibrary.com/subjects/

The subject guides use a variation of "Pirate Source" from the Joyner 
Library at East Carolina University.

http://systems.lib.ecu.edu/piratesource/source.html

Our web librarian  (Andrew Darby) adarby at ithaca.edu, has tweaked the code 
quite a bit. All of our librarians are liaisons to various departments and 
schools and are responsible for adding sources into the master database of 
electronic resources (databases, web sites) and print sources that the 
library owns. One can weight which databases are listed first.

Our electronic resources librarian puts in all the databases and their 
descriptions into the master database.

Jim Bondra
Business and Communications Librarian
Ithaca College
Ithaca, NY
607-274-1962

***************************************************************************************************************

At 01:32 PM 5/12/2005, Robert Long wrote:
>We'd like to revamp the database area of our web site and I would be
>interested in any strategies that have been developed to help users find
>the best resource to match their search needs.
>
>Are the tried and true methods of subject headings and a-z lists the
>best way to do this or is there something better?
>
>When faced with Business Source Elite or Master File Premier, what's a
>patron to do?
>
>I know that asking a librarian is always the preferred method. But if
>you had to offer a second option, what would it be?
>
>
>--
>Robert Long, Library Systems Administrator
>Talbot County Free Library
>410 822 1626 (v)
>410 820 8217 (f)

Mark

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Mark Allan
Head, Reference Unit
Porter Henderson Library
Angelo State University
ASU Station #11013
San Angelo, Texas 76909-1013
(325) 942-2511
(325) 942-2198 (FAX)
mark.allan at angelo.edu
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
I love the smell of book dust in the morning.  It smells like...... knowledge.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Any opinions given above do not necessarily reflect those of the Porter
Henderson Library, Angelo State University, the Texas State University
System, the state of Texas... 

*********************************************************************
Due to deletion of content types excluded from this list by policy,
this multipart message was reduced to a single part, and from there
to a plain text message.
*********************************************************************