WAIS gets beefed up.

Rob Harper Rob.Harper at CSC.FI
Thu Dec 24 10:21:16 EST 1992


>Xref: funic comp.infosystems.wais:745 comp.infosystems.gopher:1688
>Path: funic!news.funet.fi!sunic!mcsun!uknet!doc.ic.ac.uk!agate!ames!elroy.jpl.nasa.gov!swrinde!gatech!concert!rock!jazz!fullton
>From: fullton at jazz.concert.net (Jim Fullton)
>Newsgroups: comp.infosystems.wais,comp.infosystems.gopher
>Subject: Z39.50 Project Announcement
>Message-ID: <1992Dec23.203300.26379 at rock.concert.net>
>Date: 23 Dec 92 20:33:00 GMT
>Sender: news at rock.concert.net
>Organization: MCNC Center for Communications -- CONCERT Network
>Lines: 101


As some of you may know, the U.S. National Science Foundation has funded
the creation of the Clearinghouse for Networked Information Discovery
and Retrieval. One of the responsibilities of CNIDR is to promote the
use and development of tools for discovering and using networked
information. We have been specifically charged with the task of
providing an evolutionary path for the public-domain version of WAIS,
the Wide Area Information Server. Therefore, we have been paying close
attention to the remarks and "wish lists" provided by the network
community, and are well into a project to try to fulfill those wishes.

We are in the early beta test stages of a new generation of WAIS-like
systems that build on the heritage of WAIS as released by Thinking
Machines, but with quite a bit of extensibility and new features.
Specifically, and with the help of many others, we're creating a
Z39.50-92 (soon to be 93) server with full backwards compatibility with
current WAIS clients and indexes, features such as boolean search,
formatted searching on fielded text information, thesaurus based synonym
substitution, stemming, and a search engine integration toolkit.

While the server maintains compatibility with current WAIS clients, it
will also allow searches and retrievals from any Z39.50 client capable
of recognizing the BIB-1 or INFO-1 attribute sets. Clients developed for
this system will also be able to interact with library automation
systems within the constraints of the attribute sets. We also hope to
provide a tool to make various Z39.50 server maintenance tasks easier. 

We hope to support the following Z39.50 features: 

1) BIB-1 and dynamic INFO-1 searches

2) ES-1 element sets, which allow even more retrieval flexibility than
the current WAIS server.

3) Access and resource control - essentially, this will allow for secure
servers and servers that can accept billing information from the user.

4) EXPLAIN. Smart clients can use the EXPLAIN facility in conjunction
with their  preferred attribute set to provide customized search options
for the user.

Basic search engine features (configurable):

Synonym lookup

Stemming

Formatted field searches

Boolean search logic (to be added before release) 

We have not yet begun to create the search engine toolkit as we think we
need more user input, and the clients are rudimentary at best. We do
hope to have the server running and ready for beta-test by February,
although alpha code will be available much sooner. Clients wil take
somewhat longer as user interface development can be quite time
consuming. The entire system will, of course, be freely available for
use and modification by the network community. 

On the client side, work is progressing on clients for the Macintosh and
Unix systems, with a PC client in the wings.

So, when will this wonderful thing be ready? We actually have a server
running that answers both WAIS and Z39.50-1992 queries on the same port,
but returns results in a non-standard transfer syntax. We are in the
process of changing this to support INFO-1 generic data records, and
hope to have that completed soon. 

We have had many comments from interested folks asking us to make these
plans known to the rest of the network community, and at the risk of
announcing vaporware, are doing so. 

How can you help? We have already had quite a bit from many people, most
notably John Kunze at the University of California at Berkeley and
Clifford Lynch at UCOP. John created most of the server code and
designed the INFO-1 attribute set as well as the ES-1 element set
structures. I have made the server modifications to support free-text
searching and WAIS compatibility, as well as making the necessary
changes to the WAIS search engine. Members of Joan Gargano's staff at
the University of California at Davis are now working on X Windows
clients.

We are looking for folks willing to work on clients and run beta-test
servers with their current WAIS information bases.   Once our transfer
syntax code is firmed up, trial servers will be made available for this
purpose.

As technical manager for the Clearinghouse (as of December 1), I will
help coordinate development tasks and incorporate any of your
contributions into releases (with attribution, of course). We also have
staff who will provide documentation and support for final releases.

Your comments and suggestions are welcome, and we look forward to
hearing from you.




Jim Fullton
Clearinghouse for Networked Information Discovery and Retrieval
Phone: (919)-248-9247  E-mail:  fullton at concert.net




More information about the Bio-soft mailing list