[PLUG-TALK] [PLUG] The end of libraries

Keith Lofstrom keithl at gate.kl-ic.com
Fri Nov 14 02:35:55 UTC 2014


On Thu, Nov 13, 2014 at 01:59:15PM -0800, Mike Cherba wrote:
>    I've had good luck with google scholar.  

Google scholar /is/ good - if you are looking for a specific
paper, or something like it.  But it is not especially good
for browsing, when you don't know the jargon, or if Google's
OCR stumbles on the old text.

Some of my best "discoveries" were found by looking at old
journals ( >50 yo), and seeing what clever people did with
the inadequate technology of the time.  I frequently stumble
across great stuff while seeking other old cites.  When I
pull down a book of bound journals, I often flip through and
look at all the tables of content, not just for the article
I seek, and spend a few minutes looking at papers contemporary
to my initial goal.  They explain the mileau of discovery.

We think of ourselves as cutting edge in all regards, but I
have found that "new" technologies are often repeats of very
old stuff, sometimes replacing antique cleverness with modern
brute force.  Even if an article is online, our neologisms do
not match the neologisms of the past.  It is difficult to
search for a phrase which has disappeared, or work by authors
long since retired or dead.  

Does this matter?  I remember a "cutting edge" paper at an
integrated circuit conference decades ago, using switches and
capacitors to emulate a resistor.  A well versed attendee
pointed out that the reverse of this emulation was how Michael
Faraday came up with the idea of the resistor (he did not call
it that, of course) in the age of switches and capacitors
(Leyden jars), high voltages and very low currents.

Two years ago, I was pouring over old radar books at the MIT
Barker library and discovered a two dimensional radar hack 
that is extensible to three dimensional spherical antennas -
if lots of computation is available to simulate it.  Further
time in those old books might have discovered hand calculation
hacks that would greatly reduce the amount of computation needed. 
I suspect it is only the topography of the stacks at Barker (a
claustrophobic narrow annulus of heavy metal shelves surrounding
the cavity of the big dome, with tiny elevators) that inhibits
MIT from moving old material and shelves to a distant warehouse. 
During my unofficial visits, I had those floors to myself.

My concern is not that recent work by active researchers will
disappear soon, but that in the long span of time most of it
will, if we sequester the old work behind greatly extended
copyright and let the paper rot in distant cheap warehouses. 
Eventually, the old people who remember the older work will
retire.  Does that make the old work useless?  We can't know
until we look, and my experience is that there is are great
ideas and inspiration buried in those old papers and books.

There are also things that did NOT work then, and cannot 
work today for the same reasons.  Many research projects
and startups fail because they ignore past mistakes.

Keith

-- 
Keith Lofstrom          keithl at keithl.com



More information about the PLUG-talk mailing list