On 4/26/10, Weinheimer Jim <j.weinheimer_at_aur.edu> wrote:
>
> Alexander Johannesen wrote:
> <snip>
>
> I've yet to see how much better the library tools are. Prove me wrong.
>
> </snip>
>
> And therein lies the entire discussion. I proved the statement wrong and it
> is simply ignored, along with dismissing everything else I demonstrate. I
> desmonstrated that there is a difference between searching concepts and
> searching text, and that *people cannot search concepts in full-text tools
> such as Google. They search text. And within certain, known limits, people
> can search concepts in a library catalog.* Period. I realize that this must
> be very uncomfortable, but it cannot be denied, only ignored or to say that
> it's unnecessary.
Anyway you look at it, both opacs and Google simply match patterns of text
and return the results. To say otherwise for either is to suggest some sort
of inherent intelligence in the system itself.
Two things:
1. 80 percent of success is showing up. Google definitely shows up. It is
good enough for most people and in most situations, and it is continually
getting better all the time. Amazon too actually appears far more useful to
me than most opacs today if you want to search for a book or learn about
it. Most opacs don't shine a light to Amazon. Good enough and constantly
improving is what matters. Many librarians measure improvement of their
tools in years and decades. Google and Amazon demonstrate measurable
improvement in months.
2. In order to objectively evaluate full text tools with MARC based
catalogs, remove the human elements. You cannot fall back on the reference
librarian as an active element in a search. The tools must stand up on
their on technical merits. A knife in the hands of an experienced chef does
not make it any sharper in any objective sense. Likewise Google in the
hands of an experienced reference librarian will return different results.
If it requires an "expert" to use a tool effectively, something is wrong.
And no, users do they accept whatever Google throws at them. To make the
assumption that they do disrespects the user. Not all patrons need
handholding.
And a third thing. What a cataloger thinks is an important semantic mapping
is not necessarily what is important to the user. So all these static
semantic structures in MARC based opacs basically reflect the minds of
librarians -- not patrons. Depending on how close those two worldviews are
to eachother they either mesh well or mesh poorly. Static is an important
word here. Imagine an opac with a semantic map that actually adjusts itself
automatically to the inputs that are thrown at it over time. I'll bring
more thoughts on that idea later.
Peter Schlumpf
www.avantilibrarysystems.com
Received on Mon Apr 26 2010 - 10:22:52 EDT