On 5/2/07, Andrews, Mark J. <MarkAndrews_at_creighton.edu> wrote:
>
> You know, if I never hear anybody mention Bush, Nelson or Ranganathan
> again I won't be sorry. Haven't we had an original idea in LibraryLand
> since these guys were around? You'll notice they're all guys, too.
It's really the intersection of Moore's law with Mooer's.
Processing power has increased to the extent that we can starting testing
many of the ideas that were previously too painful and troublesome to use.
The air is heavy with the scent of new possibilities; I haven't felt
anything like this since '91, when archie, WAIS, gopher,prospero and www
burst into the light of the nets great Cambrian explosion. I was suckered
in by the siren song of search; diving into the depths of WAIS there seemed
to be so many things to try, to see what might make search results better.
After coming up with some nifty ideas, like dividing a term's weight by the
number of documents the term occurs in, I summoned up the courage to do
battle with the fearsome Aleph, guardian of the Technion libraries.
There,courtesy of Gerard Salton, I discovered just how many of the ideas I'd
had already had names. As I read, I was struck by just how powerful the
little Decstation I was using was compared to the big iron that Salton did
his work on.
WAIS came in two flavours. The one I was working with was the free version,
written for normal machines. The other version ran on Thinking Machines
CM2 supercomputers. The CM2 was capable of performing 32 GFLOPS. The CM-2
cost several million dollars. The Cell based linux box on my desk can peak
at around 200GFLOPs, at a cost of about $599 before tax.
What that power means is that we can start trying all sorts of techniques
that until know haven't been practical enough to test. We can take the LCSH
with its terms and links, feed it through clustering algorithms, and find
some of the missing structures that lead to prejudice and antipathy. We can
articulate around facets like angels around pinheads. We can build a new
bibliographic networks that recognize that megabits of bandwidth can be had
for the price of a cup of coffee; we can build new bibliographic rules that
recognize that card catalogs exist only as bookmarks; and we can build
bibliographic tools that recognize that cataloging is social.
If we've learned anything from Graphic Novels it is that with great power
comes great responsibility, and that every issue brings new challenges. In
a few years we will be reaching an inflection point in storage costs, when
it will cost more to delete things that it does to conserve them. If we
don't have the tools in place to bring a semblance of order to the flood, to
destroy what needs to be destroyed, and to at least give those who will need
it a chance at preserving the needful, we are going to swallowed up by the
digital midden pile.
OK, clock's ticking - let's move it people.
Simon
Received on Wed May 02 2007 - 10:09:25 EDT