3.705 more on NeXT (101)

Willard McCarty (MCCARTY@vm.epas.utoronto.ca)
Mon, 6 Nov 89 17:33:10 EST

Humanist Discussion Group, Vol. 3, No. 705. Monday, 6 Nov 1989.

(1) Date: Sun, 5 Nov 89 18:53:08 EST (25 lines)
From: mike@tome.media.mit.edu (Michael Hawley)
Subject: Re: 3.702 more on NeXT (57)

(2) Date: Mon, 06 Nov 89 10:36:08 -0800 (56 lines)
From: Malcolm Brown <mbb@jessica.Stanford.EDU>
Subject: using of the NeXT machine

(1) --------------------------------------------------------------------
Date: Sun, 5 Nov 89 18:53:08 EST
From: mike@tome.media.mit.edu (Michael Hawley)
Subject: Re: 3.702 more on NeXT (57)

With regard to the comments about the NeXT text-related work,
and its general applicability, it would be a mistake to think that our
indexers (or, I think, any one index/retrieve utility) will accommodate
everyone's needs. Certainly a card catalog is not the only means of
access to the contents of a library -- there's a plethora of specialized
indices, and reference/pointer sources, intended to help specialists
find what they want. Carpenter's remarks are right on the button.
We thought it was really important to advance the field towards piling
vast libraries into small boxes which can accommodate the gamut of
text-based research. I think our work is early in the larger scheme
of things -- these are the first widespread digital books which have
reached a really broad base. We don't yet search Sanskrit, or draw
Japanese very well. The major overheads in the system (including price,
which is primarily a function of the cost of present technology) are
a combination of (a) the fact that this is a young computer at the
start of its life, and (b) the fact that working with digital text
in general is hampered by a lack of availability of relatively homogeneous
technology (like the "book" interface which we've had for the last
500 years), not to mention a lack of availability of what might be
called a "critical mass" of text. This last reason is why libraries
have always been such thrilling resources.
(2) --------------------------------------------------------------67----
Date: Mon, 06 Nov 89 10:36:08 -0800
From: Malcolm Brown <mbb@jessica.Stanford.EDU>
Subject: using of the NeXT machine

I think David Carpenter's suggestion to offer ways the NeXT
system is currently being using is a good one.

We are experimenting with a text retrieval system. Since
the digital librarian (a text search program native to the NeXT)
is too underpowered, we've bypassed it. We
acquired a copy of PAT, the text search program developed
at Waterloo, and have it running on a Sun4. One our our
programming wizards has put together a front end system
that runs on the NeXT. The NeXT program "talks" to PAT
on the Sun, requesting searches and displaying the results.
I'm currently trying to enlist users here at Stanford;
if there's interest we may undertake a full development
with an eventual port to the Mac.

I think the interface is quite good (it certainly *looks* good,
at least, on the NeXT). Despite some limitations in PAT, out
NeXT front end offers nearly all the capability of WordCruncher.
We even have a graphic display for frequency charts. I'm
enouraged; if we can get a more robust search engine on the
Sun, I think we can develop a useful tool.

I'm also using the NeXT machine as a text "hub". As the
manager for our campus text scanning service, the NeXT
machine has been very useful to me for archiving and manipulating
the texts we generate (as well as those we acquire from places
such as Oxford).

I've also just completed scanning the Nietzsche corpus, and am
using standard Unix tools such as awk to assist in the post-scan
cleanup. As soon as the NeXT system supports diacritics, I'll
use it as locus for the Nietzsche master files.

I'm just beginning a review of Framemaker 2.0. My goal here
is to see how academically "relevant" Framemaker is and
whether it's worthwhile to pursue a site license agreement.

Ever since it's inception, I've been saving "clippings"
from HUMANIST into files on a Unix host. I have moved this
textbase from a Vax to a optical disk on the NeXT. I actually
do use the digital librarian to search it; for straightforward
searches the librarian is OK (at least it's more fun that
using grep...)

Malcolm Brown