Don Burn wrote:
"I've been trying it for a few days and find it frustrating versus the older
implementation. The index seems to have serious problems for instance in
the old index "B&M" found articles about the Boston and Maine, but the new
one spews 1892 references and the first one on the Boston and Maine is
The lack of the keyword list, and the only providing 10 links per page are
also annoying when you are trying to track down an article with less than
full detail. It is better than nothing, but not up to the old form (which
had the cover capability and table of contents that Tim mentioned)."
To which Tim O'Conner responded with:
"If you literally search on "B&M" (with double quotes) it gives an error.
If you search one magazine like Mainline Modeler for B&M (no quotes) it
finds a bunch of articles, and SHOWS you that B&M is a keyword for some
of the articles. But if you click on the keyword B&M it will search the
entire index again and give you the old 1,892 references as you said.
Definitely some kinks in the system! :-)"
I'd suggest those who find "kinks in the system" forward their examples to
Kalmbach for review and correction.
Specific examples of failures are much easier for the software developers to
fix if they have specific examples of requests that fail to return the
desired/expected data. Some problems like this can be fixed relatively
easily if the developers are given specific conditions that allow them to
recreate the error.
Others may not be so easy - for example, if a "B&M" keyword isn't set for an
article on the Boston and Maine, the search algorithm isn't going to find
Situations such as that would require adding the "B&M" Keyword to any
"Boston and Maine" related entries. Updating all instances of Boston and
Maine data to include "B&M" as a keyword requires identifying all the
qualifying entries in the database and tagging them. That's data
maintenance and requires discovery time plus data entry time to correct
(Though might be fairly quick to resolve using SQL or similar tools if the
database is properly designed).
If folks find "shortcomings" in the search features of the index, send those
specific examples to Kalmbach - they can't fix them if they don't know about