Reimplement TeX in Lisp? I think not...
Harvey J. Stein
Sat, 3 May 1997 17:40:52 +0300
Tim Pierce writes:
> > Date: Fri, 2 May 1997 21:10:14 +0300
> > From: "Harvey J. Stein" <email@example.com>
> > Thomas Fischbacher writes:
> > >
> > > If you have a closer look, TeX is quite horrible.
> > > I'd prefer a small LISP package providing basic typesetting functions.
> > So would I. But what am I going to do with the megabytes of *.tex
> > files?
> Port TeX. It may surprise you to learn that TeX was not developed
> on Unix, and is not wedded to the Unix paradigm. (Or it may not
> surprise you. But people seem to forget that this is an
> applications question, one which I think is largely tangential to
> the important questions surrounding the design of a new Lisp
Oy. I can't imagine that there're so many Lisp hackers around looking
for incredibly boring and tedious projects that all the tools
necessary for useing computers &/or dealing with the outside world are
going to get reimplemented. If the choice is either reimplement TeX,
groff, xv, The gimp, and xfig, or use Linux, I'd expect most people
to do the latter. Which is one of the reasons I suggest the following
steps to world domination:
Step 1. LispENV on top of Linux.
Step 2. C->Lisp compiler which allows me take an application's C
source code and compile to run inside of LispENV.
Step 3. Replace Linux with LispOS.
Step 4. Clean up LispENV <-> LispOS interface to take advantage of
the fact that LispOS is underneath instead of Linux.
As I said previously, Linux would have been still-born if it wasn't
for the mounds of free applications available for it. The same will
hold for LispOS.
So, I don't think that the applications question is tangential. I
think it's actually fundamental to the acceptance of LispOS. If the
applications aren't there, then people won't use the OS.
When people state that LispOS will succeed because Lisp is so much
better for developing applications, they're just reaffirming the fact
that it's not the OS, it's the available applications.
Philosophically, one of the fundamental reasons that people use
computers is data manipulation and transfer. I.e. - Convert a table
of numbers to a graph, stick it in a report, and email it somewhere.
And, the fundamental nature of computers is that they're programmable
machines. I.e. - an OS which only can run one application at a time
(eg DOS) is more against the fundamental nature of computers than one
which runs several simultaneously (eg unix). A computer can be a
jukebox and a mailreader at the same time. Nothing should force it to
be one or the other.
To the extent that a system doesn't support the above two points and
to the extent that there are other systems that do, then the first
system will die and the second will flourish.
So, if there are lots of apps for system A, and hardly any for system
B, but B is really easy to program, then B still doesn't have a chance
in hell until the fact that it's easy to program has lead to lots of
Similarly, MS Windows subsumed DOS because it allowed more at the same
time on the same computer. All the DOS apps were more or less
available, but MS Windows sort of allowed using more than one at a
time. Thus, it didn't interfere with people's fundamental reason for
using computers, and it was more in line with a computer's fundamental
No matter how easy it is to program under LispOS, people will still
have to talk to other computers. So, they'll still have to convert
image files btw some 50 odd formats, they'll still have to talk
TCP/IP, they'll still have to be able to display, print, and
manipulate dozens of different file formats (including postscript,
pdf, RTF, HTML, plain ascii, TeX, ...), they'll still have to connect
to HTML web servers, ftp servers, IRC chat servers, ...
If you don't provide these services, they're not going to use your
operating system. It's as simple as that. Which probably goes a long
way towards explaining why all those cool academic research operating
systems are developing Alzheimer's as the magnetic tape on which
they're stored slowly drop bits. No apps, no need.
Harvey J. Stein
Berger Financial Research