[Web4lib] Open Source Federated Searching Software - whatwe would need.

Brian Kennison brian at wcsu.edu
Fri Jun 16 11:55:45 EDT 2006


In message <45057232A4AC6444B0F9452131E7446E0391624B at exchange.win2k.tsl.state.tx.us>, "Danielle Plumer" writes:
>David,
>

>We've found Index Data to be very good at supporting open standards. I =
>don't know that I'd consider Keystone as an alternative to DSpace for a =
>true institutional repository, but it's probably on a par with =
>Greenstone and streeprints.org as a digital library option.

We use both DSpace and TKL (Index Data). In my opion TKL can
definitely be used in place of DSpace. A TKL repository can be both an OAI
service provider and can harvest OAI resources. TKL makes handling multiple
meta types (Dublin Core, Mods, Mets, you name it) within the same
repository possible while DSpace's metadata schema is fixed (which is fine
but maybe not what you're after). 

One of the things that's supposed to make an "institutional repository"
different from other web applications is the need for preservation. TKL
doesn't make any attempt at this but DSpace (as far a I can tell but I may
be missing something) doesn't really help much in this area either. Fedora
is much more concerned with managing bitstreams and I've heard of people
using TKL for managing their metadata, search and presentation layers and
using Fedora as the object store. This is something that I would like to
investigate. 

DSpace and TKL are both applications meaning they are designed for a
particular task but I tend to look at TKL (really the tools and components
that comprise TKL) more of a framework. You can use this framework to build
almost any kind of application leveraging you knowledge of how TKL works.
This is also true of DSpace in leveraging the use of Java server pages (jsp)
and tomcat but TKL seems more flexible. TKL uses xslt, php, perl but if I
want to write a function or component in python I can. Along with OAI-PMH
TKL applications are Z39.50 capable. They have a federated searching option.

While I don't claim to be any kind of expert I find the coding style of all
the TKL components to be of high quality, clear (well as clear as my little
brain can make of it) and well documented. They also have pretty extensive
documentation on their web site about how things work and how to get a site
up and running. The mailing list traffic is light but responses are very
friendly and always useful. 

I also like the Index Data business model (I wish every vendor did this).
Their software is open source which is great but they offer their services
for a fee. There are lots of reasons why you might want to employ those
services. Being selfish I'd like to see the uptake of this software grow. A
larger community means more good stuff coming from that community (which
helps me). I think TKL and Index Data in general deserve a bigger community.  

--
Brian Kennison
Systems
Haas Library
Western Connecticut State University


More information about the Web4lib mailing list