perm filename JOBS.1[LET,JMC] blob sn#871176 filedate 1989-03-14 generic text, type C, neo UTF8
COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	\input jmclet[let,jmc]
C00008 ENDMK
C⊗;
\input jmclet[let,jmc]
\jmclet
\address
Mr. Steven Jobs
NeXt Computer
3475 Deer Creek Road
Palo Alto, CA  94304
\body
Dear Mr. Jobs:

	Because of a misunderstanding with my secretary, I don't
know if this letter was mailed previously.  Anyway thanks for
the NeXt material.  If you have previous, there is no need
to read further.

	I am impressed by the attention the design of the NeXt
computer pays to its use in an academic setting.  It leads me to
hope that you may be interested in an improvement.

	The improvement is to make available arbitrary character
sets, especially mathematical character sets, in the interactive
use of the computer.  Languages like APL provide another example,
and special characters would be used more freely if the hardware
was friendly to it.  Of course, \TeX\ and other text processors
permit their use in documents, but this does not work
interactively.  You can get the symbols $≤$ and $α$ in a
document, but the file must contain {\tt\char`\\leq} and
{\tt\char`\\alpha} respectively.

	Modern bit map displays and printing devices deal nicely
with arbitrary characters.  Text editors require some special
provisions to deal with characters represented by several bytes.
However, GNU Emacs, which I understand is available on NeXt, has
these provisions.  The remaining problem is keyboard.

	Some help is coming in connection with the use of computers
for foreign languages, especially for Japanese.  However, this
doesn't quite meet the need for special mathematical characters
for academic use.  The reason is that Japanese involves a fixed
though large set of characters.  Academic use requires the
ability to use a variety of sets of characters and to make
additions.

	The keyboard problem can be handled in a variety of ways.
First of all, each character should be represented in the memory
of the machine and in transmission from the keyboard by a
certain sequence of ASCII characters.  For example, $\alpha$
might be represented by \TeX\ {\tt\char`\\alpha}.  It would be handled
as follows.

1. On an ordinary keyboard (or any not
having an $\alpha$ key) the user would type (say) {\tt\char`\\alpha}.  As he
typed, these characters would be entered, but when he hit
the space at the end, the characters would vanish and be
replaced by the $\alpha$.  The editor would have been initialized
with the list of special characters being used.

2.  A keyboard with an $\alpha$ key would emit the same sequence
of bytes when the $\alpha$ key was pressed, and the $\alpha$ would appear
on the screen immediately.

3.  A universal keyboard would have a slot for a ROM or
EPROM giving a character set.  It would also allow a character
set to be downloaded from the computer.  There would be overlays
to provide labels for the character set.  In the fanciest version
there would be LCDs on the key tops that would be set by the
ROM or downloaded from the computer.
\vfill
\eject
	This is only a sketch of what has to be done.

	However, I'm sure that computers will eventually allow
the same freedom to use arbitrary characters that a person
communicating via handwritten paper has.

	If you are interested in the possibility of enhancing
NeXt computers in this direction, I would be glad to discuss
it.  The discussion would be entirely nonconfidential on our part.

\closing
Sincerely,
John McCarthy
Professor
\endletter
\end