SWISH Wish List

CONTENTS: | Bug Fixes | Enhancements |

Please note: Many of these bug fixes have been implemented in the new version of SWISH, SWISH-Enhanced. This page remains here for historical reasons, but for the latest information please see the SWISH-Enhanced Web site.

The Simple Web Indexing System for Humans (SWISH) is a simple and yet powerful indexing system written by Kevin Hughes. The SWISH software may be freely downloaded from Enterprise Integration Technologies, which is also where you can find the documentation. Prentiss Riddle of Rice University has also gathered various SWISH bug fixes.

We use SWISH extensively here at the Digital Library SunSITE, and in so doing we have both discovered some bugs and also dreamed of some enhancements. We have recently received permission from Kevin to hack SWISH to fix bugs and add features, so long as we adhere to a few very reasonable stipulations. One such stipulation is that the result is not to be sold and will be available for distribution along with the original (which is fine by us). So we are putting out a call to all SWISH users to help us specify changes that will the most effective at improving it for our various uses. Please realize, however, that we make no promises to supply any requested bug fix or enhancement, and we reserve the right to never deliver a functioning product. That being said, we have a lot of interest in this project, as we have found SWISH to be very useful for our needs. Questions about this project may be directed to Roy Tennant.

At this stage, the fixes and enhancements are listed in no particular order or priority. If you feel strongly about a certain enhancement or fix, please let us know so we can take that into account.

Periods

A period in the middle of a word should be treated as a normal character (e.g., Z39.50), but at the end of a word (e.g., at the end of a sentence) should be ignored.

Wild-card Problems

An asterisk on the last term of a multi-term search does not work properly.

Merge

See Warren Jones' explanation of the problem.

META Tag Searching

The ability to limit searches to the contents of particular META tags. In particular, we are interested in any programming support to allow the use of Dublin Core metadata elements.

Ignore Section

The ability to specify a begin and end statement (using, for example, an HTML comment) that will cause SWISH to ignore everything between them for indexing purposes.

Thesaurus

An optional specification of a file of synonyms. This would allow a user who searches on "e-mail" to find documents using the word "email".

Phrase searching

This may be difficult.

Boolean NOT Support

Specifying words that should cause documents to drop out of the results set.

Explicit Nested Boolean

Ability to use parentheses to specify precedence.

Ordering of Search Results

Since SWISH relevance ranking is possibly bad, and in many cases is not required or desired, implement alphabetic and numeric sorts. Other kinds of sorts may be offered through a configuration file setting.

Copyright © 1997 UC Regents. All rights reserved.
Document maintained at http://sunsite.berkeley.edu/Tools/swishfix.html by the SunSITE Manager.
Last update November 19, 1997. SunSITE Manager: manager@sunsite.berkeley.edu