[Sigia-l] integrated catalogues?

[Alexander Johannesen]

Hi,

Skot Nelson <skot at penguinstorm.com> wrote:
> I always had good results from Alta Vista, and it didn't seem to me
> that Google was a quantum leap.

I think the Quantum Leap was a slow one; I too used to use all sorts
of search engines, because they all delivered various degrees of
results. I think Google won a lot by simply having the most pages
indexed, and since it also was fast, there was no reason to use
anything else.

> I'm going to use the old, somewhat dated term "fuzzy search" as an
> umbrella in which to contextualize my comments. [...]
> I suppose what I have yet to see is "fuzzy" search that truly works
> in cases where extremely high precision is a goal.

Heh, that depends on what the expression "truly works" really mean.

> On the other hand, a comment from another mailing list:
> >> I was just wondering if you had a link to the Interesting Breeze
> >> stuff. I'm not familiar with it and couldn't narrow down a google
> >> search enough to find it.
> demonstrates that Google still has shortcomings, when trying to
> perform tight searches. The searcher, in this case, had substantial
> context.

Oh, I don't think anyone would claim Google to be clever at
specialisation, and it isn't either their main game (until we see the
full extent of Google Scholar and its ilk), but even they understand
that there is fuzzy and there is special search.

I use Google for all things, and of course there are bucketloads of
stuff I can't find, especially when I'm looking for something specific
using words that are general; searching for the 'pie framework' yeilds
so many different frameworks and projects called 'pie' that it's
pretty useless. On the other hand, I use Google to search for stuff on
our own site because it does it better than our own searches (well,
for some of this stuff).

Google won't do it all, for sure, but my point was more that Google
has shown us ways of doing search better, and there is no excuse for
doing things slow if its been proven that it could be done fast. It's
just technology, whether the search is fuzzy or specific. :)

> I've been looking for details on how to check the oil level on a 1981
> Virago, and have very very little luck with Google.

   http://www.askanowner.com/qa/view.asp?guid=&sid=1&qid=17651

:)

> True. But it strikes me that to satisfy Academics, you need to do two
> things:
> a) Actually get all the information in 10 seconds instead of a
> minute, and;
> b) Convince me that you've gotten all the information, and that I
> don't need to comb that obscure DB my field has relied on for years

Unfortunately, academics learn that their world *is* these obscure
databases, and anything found elsewhere is not trustworth. Sometimes
that might even be true, but certainly not all the time. It's like
discussing the truthfulness of Wikipedia, for example.

But yes, federated searches across a wide range of DB's and pages to
make sure you've got it all covered is sometimes paramount, but this
topic dips more into political and economical limitations than
technical ones.

> The latter is harder to achieve than the former, for reasons that
> have a great deal to do with human failings rather than technology.
> It's also something that tends to transition generationally - newer
> generations trust newer technologies more naturally, having not been
> raised with The Old Ways.

Good point, although some of our most avid lovers of new technology is
some of our oldest and die-hard users. Go figure. :)

> Perhaps this is just another level to be beaten. Human searching is
> often characterized as much by targeted searching as it is by
> somewhat more random browsing...it's generally my view that we
> haven't solved the browsability issue in software yet.

True; I try to create systems that do 50% of each, usually by
clustering, semantic analysis, using Topic Maps, etc, etc. To do any
of this in a meaningful way you have got to get faster; sitting and
waiting for results over a log period of time is probably fine if
there is a guarantee that the results are going to be perfect, but
even the best specialisation tools out there can't provide such a
guarantee, and hence the human tendency to 'move on to faster things',
even in academia.

> > In fact, my service
> > described above *is* an academic search, and 10 seconds for a search
> > was borderline enough for us to try and be smarter and faster about
> > it. And seriously, it isn't that hard.
>
> Hmm. Is this something that a Canadian could dive into, or does it
> require some sort of restricted access? I'd love to poke around.

Sure, although it hasn't been officially launched yet. The old
"indexed search" slow version is at ;

   http://www.nla.gov.au/apps/randr

Some bugs and worts intact. And don't treat it like Google; it doesn't
like it ... too much. :) The idea here is searching and browsing
topics we know our patrons ask about (search logs, oral interaction
logs, etc) which we have resources on. The thesaurus (APAIS) is
perfect for finding words we do further searching on, so if the main
word *and* an APAIS related term is found in a resource, it ranks
higher than a resource that only had the main word in it. And 100
variations around that. :) And all grayed out resources are not
available unless you are a) logged in, or b) on-site.

> > Libraries have a long tradition of designing things that are hard to
> > use, cryptic to understand and generally produce sub-par results.
>
> Hee hee.
>
> And yet, I love walking through libraries. Robarts Library at the
> University of Toronto was a particularly impressive temple to the
> printed word.

Libraries are still pretty cool buildings, but their use is definitly
changing. I'm wondering about what they will become, to be perfectly
honest.


Alex
--
"Ultimately, all things are known because you want to believe you know."
                                                         - Frank Herbert
__ http://shelter.nu/ __________________________________________________