Newsgroups: comp.databases,comp.lsi,comp.parallel.pvm,comp.parallel.mpi,comp.org.acm,comp.org.ieee,comp.protocols.misc,comp.realtime,comp.arch,comp.software-eng,comp.sys.super,comp.theory,comp.dsp,sci.math
From: ecale@cray.com (David Ecale)
Subject: Re: Publishing Scholarly Work on the Web -- opinions?
Organization: Cray Research, Inc.
Date: 25 Sep 1996 13:20:27 GMT
Message-ID: <52bbir$fse@walter.cray.com>

Please note: I have no objection to the premise of the title topic of this
thread.  I am only attempting to comment on some of the mis-conceptions
that I see with regard to storage of the articles themselves:

First: ale2@psu.edu writes:
<> ecale@cray.com (David Ecale) writes:
<> > Price isn't the question.  It's durability of the media.
<> > 1) The change in technology is such that data stored on media only 10 years
<> > ago is completely unreadable, well almost!
<> 
<> Shouldn't optical storage of information (cd's ect.) last in principle
<> for 1000's of years even if the machines which read the information
<> don't?
<> 
Yes, I believe that with proper storage that this may be the case.  Now,
the question that I was attempting to ask was, "Will there be equipment
available 10 years into the future to read these disks?"  If you check my
original posting, I mentioned that I had an 8"-SS-Floppy ca. 1984.  I can
no longer read this disk.  Not because of the fact that the data has
perished, but because the machines that read it have!  Let's not count out
the fact that the Word Processor used is also gone & I would have to use
cryptoanalysis to decipher the word processor codes if I didn't have a
plain-text version.  Web pages are more readable than this, but I'd bet
money that it'll be a heroic effort to find the hardware in 10 years to
read the files from storage media used today.  Let's not even think about
1000s of years.  (PS.  The only media that I know about that has survived
1000s of years is stone (the Rosetta Stone is an example) and clay (baked
in the case of Babylonian Cuniform and unbaked in the case of Minoan
Linear B).  So, I could get my data off of that 8"-SS-Floppy, but it would
take the efforts of a small museum's staff to accomplish the task. Now
think of all of the other *dead* storage media invented for computers in
the past 50 years!  (60-col cards, 80-col cards, Magnetic strips in clusters,
paper & acitate tape, drums, myriads of tape & disk types, etc.)

Second: djb@koobera.math.uic.edu writes:
<> In article <5295oe$qj2@walter.cray.com>, David Ecale <ecale@cray.com> wrote:
<> > 1) The change in technology is such that data stored on media only 10 years
<> > ago is completely unreadable, well almost!
<> 
<> Who cares? Keep data online, all the time. Replicate data from older
<> media to newer media. What we want to keep are the bits, not the disks.
<>
Well!  Have you considered the continuing cost of this method?  Especially
considering the ever increasing volume of information & data?  Think it
over for a bit.  I suspect that, while the density of media continues to
increase, and therefore the ease of storage due to reduced raw number of
storage envelopes (disks, tapes, etc.), the overall increase of information
will overwhelm even this efficiency.  Now take into acount the continuing
obsolesence of data formats & storage media types and the requirements of
transferring the data every (I'll be cruel here) 4 years.  (As an aside, I
have used 5 different word processing systems in the past 11 years.  One of
my texts has been converted & re-formatted for each one's unique form of
markup.  HTML has been the last, to date.)  Now address the cost of keeping
that data center with it's disk farm active.  Please don't say, "Well, I'm
just storing the stuff on-line on my workstation."  You need admin
assistance, power, new hardware every 4 ~ 5 years (which ain't cheep,
either)....  

"Replicate data..."  Hmmm...  That function was peformed by monks in the
middle ages.  They would copy texts from an original to a new media
(replacing age damaged vellum & parchment).  There was a fatel flaw in
their system that we still have.  In most cases they did not have the skill
to identify what was worth saving & what wasn't worth saving.  So they
attempted to save everything.  This caused critical texts to be lost due to
deterioration of the media as there weren't enough monks to keep up.  An
especially painful example of the loss of critical texts (and the saving
of others) was the Cotton Fire in the 18th cent. (While this is not the
Middle Ages, the example still stands.)  The Cotton family in England had
an extensive collection of unique texts.  The building burned to the ground!
Nearly everything was destroyed.  One of the critical texts that was rescued
was a small book containing the only extant copies if "The Pearl" and
"Sir Gawain & The Green Knight".  Your on-line data *is* backed up & stored
off-site on a regular basis, I hope.  (At additional cost, I regret to
say....)

Anyway, I digress.  Optical diska may last for an extremely long time &
on-line maint. of data may work out well.  The problem is that the media
becomes obsolete at a ferocious rate.  The data formats become obsolete at
an even more ferocious rate!  The cost to keep up is enormous.

One final note:  I have a copy of that Cotton text mentioned above.  I
purchased it on microfilm from the British Library.  (Anybody got a
microfilm reader?)


-- 
David Ecale
ecale@cray.com			Work = 612-683-3844 // 800-BUG-CRAY x33844
http://wwwsdiv.cray.com/~ecale	Beep = 612-637-0873
Will hack UNIX(TM) for food!

