[1235] in Public-Access_Computer_Systems_Forum
Re: Arms Clarification
daemon@ATHENA.MIT.EDU ((John Franks))
Wed Sep 16 13:26:20 1992
Date: Wed, 16 Sep 1992 12:18:17 CDT
Reply-To: Public-Access Computer Systems Forum <PACS-L%UHUPVM1.BITNET@mitvma.mit.edu>
From: "(John Franks)" <john@hopf.math.nwu.edu>
To: Multiple recipients of list PACS-L <PACS-L%UHUPVM1.BITNET@mitvma.mit.edu>
----------------------------Original message----------------------------
In article <9209141803.AA02722@wulib-jpb.wustl.edu.noname> "(Paul Buchanan)"
<buchanan@wulib-jpb.wustl.edu> writes:
>I think it's easy to over-draw Arms' analogy between libraries and computing
>technology. The tensions which have occurred over the last 10 or 15 years in
>academic computing have arisen primarily because of a change in technology: it
>is now technically possible to distribute computing technology whereas it was
>formerly difficult.
...
>
>In the larger view, this isn't quite the library paradigm. In fact, it has
>always been possible to distribute and decentralize library service and there
>is strong historical precedent for doing so.
The analogy is not perfect but has some merit. The reason we have
centralized libraries is that a single individual cannot afford to own
all the books she may need to access. As a result groups, usually
institutions or communities, pool their resources to acquire a
substantial collection and share its use. The physical nature of books
tends to make these collections be "geographically centralized", i.e.
libraries tend to be in a single physical location since the
alternative is diminished ease of use by patrons and/or time delay in
access.
Historically computing was centralized for similar reasons. The advent
of small cheap computers has brought radical decentralization. Textual
information is currently undergoing a technological change every bit as
dramatic as the change in computing. It is an interesting question to
what extent this will result in decentralization of textual (or other)
information.
There are certainly trends toward decentralization, for example, the
near explosive growth in gopher servers around the world linked in an
enormous but disorganized conglomeration of information. On the other
side electronic publishers are trying very hard to keep a rigid control
over the distribution of their products -- even to such an extent, in
my opinion, that they sometimes cripple their sales. Insofar as
publishers retain this control there will continue to be economic
forces to centralize information and share it among a "community."
However, with electronic information there is no longer any reason for this
community to be geographically centralized. We might therefore see
electronic libraries which are organizationally centralized, but
geographically decentralized.
For non-commercial or public domain information there seems to be
little reason to expect even organizational centralization. The
loosely linked, world-wide, and rapidly growing "gopherspace" gives us
some idea of what truly decentralized information bases might be like.
In any case it seems likely that we will see dramatic growth in
decentralized information bases.
What isn't yet clear is how one will be able to find anything in such
an enormous maze of information. People in the gopher
community are currently trying to address this problem. There have been
suggestions of using traditional library classification schemes to make
a catalog, but such a project brings Sisyphus to mind. The growth of
information will certainly outpace any catalog. The only hope is for a
solution based on the same technology driving the change in
the information base. I don't know what such a solution would be, but
it is instructive to note how the technology is changing "cataloging"
on a personal scale. For example, I know an individual who has a large
electronic correspondence and a large number of documents which he uses
in his work. Rather than file this material in an traditional or
organized manner, he dumps it all in a single large directory on
his NeXT and builds an index of everything. Whenever he wants a
particular document he uses search software (similar to WAIS) to find
it. Extrapolating this procedure to the internet one arrives at
something like the Archie service. What we don't yet know is how well
such schemes "scale up". Are they feasible for indexing not only
titles but also content? Will they still be feasible when the internet
is 100 times its current size?
There is a great deal here we can't predict, but I, for one, think it will
be great fun to watch!
John Franks Dept of Math. Northwestern University
john@math.nwu.edu
