11.0641 Re: Where does it go? Where has it gone?

Humanist Discussion Group (humanist@kcl.ac.uk)
Sat, 14 Mar 1998 19:07:34 +0000 (GMT)

Humanist Discussion Group, Vol. 11, No. 641.
Centre for Computing in the Humanities, King's College London

[1] From: "Malcolm Hayward, English, IUP, Indiana PA 15705" (15)
Subject: Re: 11.0636 Re: Where does it go? Where has it gone?

[2] From: "David L. Gants" <dgants@english.uga.edu> (36)
Subject: Re: 11.0634 Where does it go? Where has it gone?

Date: Thu, 12 Mar 1998 08:18:51 -0500 (EST)
From: "Malcolm Hayward, English, IUP, Indiana PA 15705"
Subject: Re: 11.0636 Re: Where does it go? Where has it gone?

I think Christoph Eyrich has it just right. The upgrades
are forced on us not by features but by the need to have
all the various systems compatible. But his response and
that of Patricia Galloway suggest an alternative to Jim
Marchand's global upgrading: archiving in ASCII--a step up
from paper copies for re-scanning, but still always readable.
Well, I guess this is part of the TEI plan. But I had not
considered adopting it for my own files.

And an aside on the subject, since most upgrades affect merely
the formatting of documents (new fonts, etc.) I have a certain
nostalgia for typing in <unl -5> at the beginning of a paragraph
and knowing that a tab will appear in the printing, in the
same way that first writing html texts allowed you to imagine
what would appear on the screen. There was a kind of pleasure
in that ...

Malcolm Hayward

Date: Sat, 14 Mar 1998 14:01:17 -0500 (EST)
From: "David L. Gants" <dgants@english.uga.edu>
Subject: Re: 11.0634 Where does it go? Where has it gone?

From: "Gregory J. Murphy" <rejek@phoenix.Princeton.EDU>

On Thu, 12 Mar 1998, Jim Marchand wrote:

> Just some animadversions on the reality we live in, brought on by facing over
> 1000 5 1/4 diskettes which I have to convert; I have only one computer left
> which reads them. In fact, as computers are `improved', we humanist users,
> who have little control over what we use, are faced over and over with the
> necessity of converting. Do you remember Vari-Typers? IBM Selectrics? Edge-
> punched cards? Paper tape? What do you do with 5 or 6 thousand punched cards?

All media have a finite lifetime; magnetic and optical media may
seem short-lived when compared in absolute terms to (some) forms of
organic material, but think of it in relative terms: how costly is it to
copy data periodically, where the period is determined by the expected
lifetime of the media? Does that cost justify the media's longevity? There
are many ways of measuring cost. One might be the total time it takes to
transfer the data.

I am told that most of the books printed nowadays won't last very
long, so let's compare with parchment. I am told that a good compositor
can lay down on the order of 1000 characters per hour. Let's take an
upper bound of a super-human compositor who never stops for food or sleep,
and can lay out at a rate of 2000 chars per hour. It will take him 1,440
hours to compose a text of 1.44 x 10^6 characters, or about what will fill
a 1.44 MB floppy. I am told that parchment lasts from 2,000 to 4,000
years. Let's take an upper bound of 5,000 years, or 4.37 x 10^7 hours.
That gives a ratio of time of production / time of duration of 3.30 x
10^-5. The slowest drive in my office will copy that much data onto a
floppy diskette in about 1.5 minutes. Say we are coppying from diskette
to diskette, which will take 3 minutes plus a minute or so to get the
label on (I always get them on crooked). If we have to perform this once
a year, either to guard against slow decay, or, as Mr. Marchand
laments, to keep up with the latest technological migration, then we are
still operating within a expense/benefit ratio of 7.6 x 10^-6, a factor of
ten better than print on parchment.

This is, I know, a gross over-simplification, and I may have
picked an invalid comparison (type setting isn't exactly the same thing
as data copying), but I just wanted to point out that, as in all things,
you have to think in terms of economies of scale.

- Gregory Murphy

Humanist Discussion Group
Information at <http://www.kcl.ac.uk/humanities/cch/humanist/>