NeXT, cont. (138)

Willard McCarty (MCCARTY@VM.EPAS.UTORONTO.CA)
Mon, 17 Apr 89 18:47:57 EDT


Humanist Mailing List, Vol. 2, No. 850. Monday, 17 Apr 1989.


(1) Date: Mon, 17 Apr 89 04:51:43 EDT (91 lines)
From: mike@tome.media.mit.edu (Michael Hawley)
Subject: NeXT comments

(2) Date: Mon, 17 Apr 89 11:29:02 EDT (27 lines)
From: Geoff Rockwell <rockwell@utorgpu>
Subject: Re: more on the NeXT (48)

(1) --------------------------------------------------------------------
Date: Mon, 17 Apr 89 04:51:43 EDT
From: mike@tome.media.mit.edu (Michael Hawley)
Subject: NeXT comments

A brief reply to comments offered by Malcolm Brown and James Coombs.

Keep in mind that release .9 of the software is intended for
"aggressive early users." Now, nobody in the real world has even seen .9 yet;
people have seen release .8, which was for hard-core *developers*.
Suffice it to say, the machine is just a little baby computer at the
beginning of its life, and by the time a polished release 1.0 rolls out
(expected late June/July) the novice perception will be quite different...

Regarding Malcolm's comments --

I have assumed all along that the feature of primary interest for
humanists is the digital librarian, a text retrieval program that is
native to the NeXT operating system. In this current form (release .8)
there's not much you can do with the librarian, other than search
some predefined "targets," such as the on-line documentation, the
Shaekspeare files and the Oxford quotations....

Most critically, I think we must look past the glitz and the well-designed
interface to see if the machine will do the work we require. It's very
nice to include the complete works of shakespeare, but if the librarian
only performs simple searches then there's not that much that differentiates
the next machine from a sun or a Vax (at least from the standpoint of
text analysis). Nor is there much to induce one to abandon the machines
running programs such as WordCruncher, MOCP, TACT, etc....

Hm, I should think that many features of the machine will excite humanists,
especially if they care at all about anything besides textual analysis.
TeX and LaTex, Mathematica, bundled word processors and graphics tools,
multimedia mailers, audio capability, and unprecedented integration and
ease of implementation of applications all add up to something pretty
synergistic.

The fact that we went to the trouble of bundling more than 50Mb of text
(about 1/4 of the disk) right out of the gate says something very profound
about what NeXT intends to do over time. It's an amazing thing for a computer
company to do. The Librarian application per se is, as Malcolm points out,
relatively simple minded: it is a channel for searching through large areas
in a content-oriented way. Yes, the .9 version supports grep-style searching
as well as index building, however it does not support things like tagged
file formats, nor does it yet have adequate support for multilingual character
sets (although diacritics are coming) or a number of the other things that
will be required in the fullness of time. It goes without saying that,
to paraphrase Varese, just because you've invented the automobile, doesn't
mean I should go and shoot my horse. But many of those other programs
will port trivially to the NeXT (and probably run a lot faster).

As far as tinkering with code is concerned, this can probably be made to
happen. One of the trickiest problems with delivering systems solutions
is that as soon as we release a spec, it tends to freeze hard. Well,
if the searching and text format specifications were released, and spawned
a cadre of great third party applications, but didn't scale up to Kanji and
Turkish, there would be great backward-compatibility problems. The other
problem I think Malcom and others indirectly sense is that a great text
application is necessarily a synthesis of a lot of key technology -- O.S.,
graphics, searching, doc. display and interchage, etc, etc. When you
build a new computer from the ground up, synthesizing full and mature
applications takes time due to the multitude of labor pains. This is
not an excuse at all, merely a fact. It is also the reason that every
scholar who ever uses a NeXT will at any time be able to select a word,
push a button, and see a Webster's definition of it (say). Bringing this
kind of synergy to people is something we thought was terribly important.

My personal goal is, in 10 years, to put a box on the shelf of every scholar
that
contains 100Gb (that's 100,000 books) that can be combed and analyzed in a
variety of ways. This is well within the realm of technology.
To help the world get there, hopefully lots of humanists
and others will chip in and evolve standards, and short term, I'm sure this
will include collaborative efforts between NeXT and scholars who want to
write code... Please, keep beating up NeXT to help us cultivate the field.
And remember, it really is just a baby computer whose formative years
are only beginning. Your input is incredibly important!

cheers,
Michael Hawley
mike@media-lab.mit.edu
mjh@next.com

[I asked Mike Hawley to describe his connection to the NeXT, and the
following is what he said. --W.M.]

I am a software engineer and head of the Digital Library group.
I implemented most of the digital books on the NeXT machine
and helped Steve negotiate all the arrangements.
I do maintain a dual citizenship -- doctoral work at MIT as well.

(2) --------------------------------------------------------------32----
Date: Mon, 17 Apr 89 11:29:02 EDT
From: Geoff Rockwell <rockwell@utorgpu>
Subject: Re: more on the NeXT (48)

One interesting feature of the NeXT interface is its scroll bar. The size
of the scroll box in relation to the scroll bar depends on what proportion
of the text you are seeing. In other words, if you are seeing one of two
pages the scoll box will be larger than if you see one of ten. The result
is an approximate sense of the size of the document opened which Mac scoll
bars do not give unless you start scrolling. This feature was borrowed from
GEM I am told.

NeXT is supposed to become more Mac-like. Both buttons will do the same
thing unless you go into a "control panel" and reset the second button.

The main problem now is that the NeXT interface does not seem complete.
There is no tool for adding and removing fonts despite the advantage of using
display PostScript. To find files among the thousands around you have to
open up a VT100 emulation and use the traditional commands like "find..."
You cannot, yet, be ignorant of UNIX when using the NeXT. That is not to say
that access to the traditional tools is not usefull. The NeXT promises
to give us the best of both but does not yet deliver. It is an experiment
in the cohabitation of command and point-and-click interfaces which will
probably shape the interfaces to come.

Geoffrey Rockwell
rockwell@gpu.utcs.utoronto.ca