Verity and default delimiters

[Thomas Dowling]

 ----
From: Walt Howe <walthowe at delphi.com>
To: Multiple recipients of list <web4lib at library.berkeley.edu>
Date: Tuesday, June 17, 1997 12:29 AM
Subject: Re: Verity and default delimiters

>Janet Kaul wrote:
>
>> Our company has just purchased Verity search systems (we have been
>> using Fulcrum).
>>

>> Has anyone else using Verity done usability surveys to indicate
>> if the comma is at all intuitive acceptable? Any stories I can take
>> to my management to convince them a change in delimiters is necessary?
>>
>> Or maybe to show me that it isn't?
>
>I first used comma delimiters with DBase2 on an Apple 2 under CP/M. I've
>used it ever since with various other software and platforms, and it
>certainly is the most common form of delimiting. I've used other forms,
>too, but it is really the least common denominator of delimitation. I
>don't know if you'd call it intuitive or not, but nothing is more
>common.
>

Mr. Apple, meet Mr. Orange.  Walt seems to be talking about comma-delimited
fields in databases, which are common; Janet is talking about
comma-delimited search strings for a search engine, which aren't so common
("FIND: apples, oranges" vs. "FIND: apples and oranges" or "FIND: apples or
oranges").

We recently tested a large bibliographic database built on Verity, and no
one here could stand the comma-delimited search strings.  I admit that we
did not perform any empirical tests (and neither did the vendor), but the
consensus among librarians evaluating the product was that we wanted spaces
to denote adjacency and Boolean operators to handle Boolean operations;
none of us could figure out exactly what the commas meant; and we got less
and less enthusiastic about relevance ranking as we looked at the product. 
Obviously different criteria will prevail for different kinds of
databases.

The solution we were pushing for, until the vendor decided to build the
next version on a different search engine, was to have a CGI front end that
would accept search strings the way we wanted to enter them and reformulate
them as necessary for Verity.  Now we'll never know how well that would
have worked.


Thomas Dowling
OhioLINK - Ohio Library and Information Network
tdowling at ohiolink.edu