Hi list,
Well, I received several very helpful responses, and about twice as many requests for those responses, which indicates to me that I am not alone in my quest for the perfect search engine. Here are the responses I received, edited and summarized:
_____________________________________________________________________
Access to Archives (http://www.a2a.org.uk/) uses TEXTML Server - see
http://www.ixiasoft.com/default.asp?xml=/xmldocs/webpages/textml-server.xml
______________________________________________________________________
I've been working with a programmer at OhioLINK, testing DLXS and PLEADE with EAD finding aids, as part of a project to establish an EAD repository for institutions in the state of Ohio. Both products are continuously in development, so many of the things mentioned below may change.
These sites give more information about the technical requirements for the two search engines, along with licensing information for DLXS:
http://www.dlxs.org/
http://www.pleade.org/en/
DLXS is a for-purchase product, with license restrictions. OhioLINK already uses it for e-books, the application for which it was originally designed. DLXS has produced a finding aids class, with plans to develop it further. Its delivery of EAD was successful and the presentation of finding aids was clear, but currently, the ability to change and customize headings and other display elements is very limited. Changes like these have to be made at a low level of code, which would cause big problems each time there is a new release. Since our focus is on designing a format that can be agreed upon by various institutions from across the state, flexibility in presentation is essential. Fortunately, DLXS's next release will incorporate XSLT. They have emphasized that this will enable local users to customize displays much more easily, and without having to edit CGI code.
PLEADE has the advantages of being free, open-source, and designed specifically for EAD. The OhioLINK programmer has found it very easy to customize display elements. It provides detailed results, focusing on each component record, with navigational features that establish context, moving from the specific component level up through to the collection <unittitle>. I have found that some users think the component-level records are too isolated from the document as a whole, while others like the specificity, and find it easy to navigate to the larger document (none of the users were archivists, incidentally, so neither group had any privileged knowledge of finding aids). The PLEADE developers are working on a module that will allow users to present context in different ways, so we may try that once it's ready. We've found that our current PLEADE setup is slow in loading finding aids with more than 1,000 <c>-level records. The OhioLINK programmer is trying a couple of different
methods of speeding things up. PLEADE was developed in France; being able to read French is a big help in following the documentation (some, but not all, of which has been translated into English) and the ongoing discussions of its development and implementation.
-Amy McCrory, OSU
____________________________________________________________________________________________
Duke University has been using DynaWeb since 1995, and due to the fact that DynaWeb is no longer being supported, Duke staff are looking at several other software tools, including:
DLXS: http://www.dlxs.org/
PLEADE: http://www.pleade.org/
LEADERS: http://www.ucl.ac.uk/leaders-project/
Archivists Toolkit: http://euterpe.bobst.nyu.edu/toolkit/
Lucene: http://lucene.apache.org/java/docs/
Xindice: http://xml.apache.org/xindice/
SleepyCat: http://www.sleepycat.com/products/xml.shtml
Cheshire: http://www.cheshire3.org/
Swish-e: http://swish-e.org/
Harvest: http://harvest.sourceforge.net/harvest/doc/index.html
eXist: http://exist.sourceforge.net/
Greenstone: http://www.greenstone.org/
XTF: http://www.cdlib.org/inside/projects/xtf/
Some of these seem to be on the verge of obsolescence themselves, others are
still in development, and others seem to only provide the searching part of the solution
without having a sophisticated way to display search results to browsers or permit simple browsing.
XTF from the California Digital Library is looking promising because it appears to be a comprehensive solution, has been implemented successfully at CDL, is open source, and has a very well thought out architecture.
_________________________________________________________________________________________
Thanks to all of those who responded! If anyone else on the list has comments about any of these search tools, feel free to share them with me. At our institution, we are looking for an inexpensive solution. We just found out we're getting WebFeat for free through a cooperative network, and we'll look into using that too. WebFeat is a federated search engine from Dynix. I have no experience with federated search engines so I have no idea how well it will work yet.
-Alison Hinderliter
************************************************
Alison Hinderliter - [log in to unmask]
Project Archivist
The Newberry Library
60 W. Walton St.
Chicago, IL 60610-7324
(312) 255-3694
|