/usr/lib/info -- hacker/librarian haven
Front Page News Features
Software Events Standards
Ask Anything Opinion Diaries
Reviews . MLP . Everything
Adding Z39.50 to non-Z39.50 databases (that you don't own)

By jaf, Section Ask Anything
Posted on Thu May 1st, 2003 at 09:10:51 AM EST
In the never-ending quest for a one-stop search shop, the information technology community has taken a variety of approaches. One approach is to harvest metadata all into a central repository - such as in the use of OAI-PMH. The other major approach, federated searching, is mainly accomplished in libraries through the Z39.50 protocol. However, many publishers and vendors do not provide Z39.50 access to their databases. The potential is there for libraries to build their own Z39.50 access to these resources, however, through tools such as SimpleServer, which, by hacking through the web interface to a database, can provide a standard Z39.50 interface into a normally non-Z39.50 enabled resource. This is being done now on the Scholar's Portal project, and I imagine this sort of approach will prove popular in the future.

So, this begs the question: Is this legal? Most of the licenses I have seen don't directly address a library's ability to 'hack' the web interface for the purpose of integrating a database into a centralized search tool. Many licenses do have clauses about automated downloading of records, and some Term of Service restrict the repurposing of search results. I'm not sure what most vendors think of others adding a Z39.50 interface onto their databases - should they be concerned, or is this actually of benefit to the vendors? What would be the best approach so that a library ensures it doesn't violate its license with the vendor? Finally, is anyone interested in building (via SimpleServer) a collection of Z39.50 scripts for various high-use databases?

 

< BatchCat and Wine (2 comments) | Koha and the Trophees du Libre (2 comments) >

Menu
submit story
create account
faq
search
recommended reading
editorial guide
masthead

Login
Make a new account
Username:
Password:

Related Links
SimpleServer
Scholar's Portal
Term of Service
More on
Also by jaf

View: Display: Sort:
Adding Z39.50 to non-Z39.50 databases (that you don't own) | 3 comments (3 topical, 0 editorial, 0 pending) | Post A Comment
Issues and possibilities (none / 0) (#1)
by Sebastian on Thu May 1st, 2003 at 03:38:54 PM EST
(User Info)

To the question of information vendor licenses, I think the only safe way to ensure that you're not stepping on anyone's toes is to ask the owner of the site, unless they explicitly allow the practice in their license statement. In most cases, you will probably have an easier time explaining that you want to use their resource in a specific web-based application. If you tell them that you want to make a Z39.50 server for their site so that anyone can search their contents without going through their interface, this will seriously freak out the majority of commercial information system operators and a discouragingly large proportion of supposedly "public" services. Fortunately, some are more open-minded.

In response to your last question, I would be interested in discussing ways to maintain collections of 'gateway scripts' for interesting databases. Since gateways to pure HTML-based user interfaces (screen-scrapers) are very sensitive to changes in the user interfaces that they talk to, I think that an essential part of such a collection has to be a robust maintenance scheme to allow script updates to flow out to the applications that use them.

I have speculated that a purpose-specific, formalized language describing HTML interfaces might be preferable to the more-or-less loose constructions people make with, say, SimpleServer (Perl, in that case). The resource description used in Apple's Sherlock seems to work pretty well up to a point, but I haven't looked at it closely enough to see if it gives enough control to build even a simple Z39.50 server on top. Unfortunately, my personal impression is that as websites get more and more creative with their use of cookies, frames, Javascript, etc., so the gateway scripts also have to become increasingly creative to keep up.

Cheers,

--Sebastian, Index Data

[ Reply to This ]


Z39.50 front-end for Relational Databases (none / 0) (#2)
by Mike on Fri May 2nd, 2003 at 07:47:30 AM EST
(User Info) http://www.miketaylor.org.uk/

The original article finished by asking:
Finally, is anyone interested in building (via SimpleServer) a collection of Z39.50 scripts for various high-use databases?
I built pretty much exactly that about a year ago: it's called zSQLgate and its website is at sql.z3950.org

The idea is that, by using Perl's DBI interface, it should run against pretty much any widely-used database system. I've used it mostly against PostgreSQL and MySQL so far, plus the wonderful DBD::CSV back-end that uses CSV files as a noddy relational database! So if you ever wanted to serve records from a CSV file via Z39.50, this may be the easiest way!

On this forum, I am a bit embarrassed to admit that, unlike all the other software I've put out there, zSQLgate is not Free (neither as beer nor as speech). It's free to download and evaluate (no registration process or similar) but there's a licence fee for deployment. I couldn't find anything in the FAQ or Editorial guidelines saying that this is a Free Software only zone, but it seems to be very much in the FS spirit.

Anyway, if this is useful to anyone, great; if not, pretend I never spoke.

[ Reply to This ]


Adding Z39.50 to non-Z39.50 databases (that you don't own) | 3 comments (3 topical, 0 editorial, 0 pending) | Post A Comment
View: Display: Sort:

Powered by Scoop
All trademarks and copyrights on this page are owned by their respective companies. Comments are owned by the Poster. The Rest 2002 The Management

front page | submit story | create account | faq | search