[WEB4LIB] RE: Two steps forward, three steps back

[Karen Harker]

While our ILS is Sirsi, we have implemented Endeavour's ENCompass for
Resource Access system, which parses queries the same way as in Voyager.
 Because of this, we, too, have implemented our own Javascript which
"normalizes" the keyword using these steps, not necessarily in this
order:
1) Removes any "and" and "or" from the query
2) Removes punctuation
3) If user enters keyword within quotes, the quotes are removed and the
type of search is changed to "As a phrase".
4) If user had entered "or" in the query, the query type is changed to
"Any"

This eliminates about 75% of user-interface errors.  I really wish
Endeavor would automatically include such a tool in their systems, at
least as an option.




Karen R. Harker, MLS
UT Southwestern Medical Library
5323 Harry Hines Blvd.
Dallas, TX  75390-9049
214-648-8946
http://www.utsouthwestern.edu/library/

>>> Andrew Darby <adarby at ithaca.edu> 12/22/2004 8:59:11 AM >>>
For what it's worth, we use Endeavour, and the reference librarians
make 
the same complaints as Stacy.  Most grating is the fact that the 
"command" keyword search defaults to OR, and Endeavour doesn't seem to

allow one to switch it to a default AND.  In this googly world, you'd 
think that would be a reasonable option.  A couple of libraries (UCLA,

Getty Research Library) have written Javascripts that automatically 
populate the search string with an AND.

Andrew Darby
Web Services Librarian
Ithaca College Library
http://www.ithaca.edu/library/ 

Jimm Wetherbee wrote:
> Bill,
> 
> [you wrote]
> 
> | 
> | If your consortium had chosen ALEPH, you would not be having the
> | problems you see.  ALPEH has its peculiarities but its searching
and
> | indexing for keyword is quite good.
> [--jimm replies] 
> 
> I tried all but one of Stacy's queries on our Endeavor system in the
same
> manner as she describe and was quite successful.  The only reason I
didn't
> attempt one (the exact title search) is because it is an option we
> deliberately left off.  I was also somewhat confused by the comments
on
> keyword searching.  Endeavor makes a distinction between a command
search
> (which uses Boolean logic, and is aimed at those of us who still
remember
> running BRS and Dialog from a terminal) and a keyword search (which
does
> not).  It is possible that her local system admin relabeled the
command
> search as "keyword," but if so, that is hardly a vendor issue at
all.
> 
> --jimm
> | 
> | Bill Drew
> | drewwe at morrisville.edu 
> | 
> | > -----Original Message-----
> | > From: web4lib at webjunction.org 
> | > [mailto:web4lib at webjunction.org] On Behalf Of Stacy Pober
> | > Sent: Tuesday, December 21, 2004 5:40 PM
> | > To: Multiple recipients of list
> | > Subject: [WEB4LIB] Two steps forward, three steps back
> | >
> | >
> | > Here are some search examples that fail and what Voyager wants to
see:
> | >
> | >   Failed basic "Find Keyword" search:  death of a salesman
> | >   Working version: death and of and a and salesman
> | >    (User MUST use boolean connectors even though it is explicitly
a
> | >    keyword search)
> | >
> | >   Failed advanced "Find in Keyword Anywhere" search: war and
peace
> | >   Working version: war peace
> | >     (Using boolean connectors in this search ensures the search
> | >      will fail)
> | >
> | >   Failed advanced "Find in Title" search: lion, the witch,
> | > and the wardrobe
> | >   Working version: lion, the witch, the wardrobe
> | >    (Again, system chokes when finding unexpected boolean "and")
> | >
> | >   Failed advanced "Find in Title": kings and desperate men
> | >   and
> | >   Failed basic "Find in Title, Exact": kings and desperate men
> | >   Working version: kings & desperate men
> | >     (Users must use ampersand if it is part of the title.
> | > System does not
> | >     know how to alias "and" and "&".
> | 
> | 
> | ---
> | [This E-mail scanned for viruses by Declude Virus]
> 
> 
> 
> 
>
*********************************************************************
> Due to deletion of content types excluded from this list by policy,
> this multipart message was reduced to a single part, and from there
> to a plain text message.
>
*********************************************************************
>