MCL essential to Macintosh

[Mark.Perlin@J.GP.CS.CMU.EDU]

Rick,

I appreciate Apple's concern with the apparent cost of the Macintosh Common
LISP (MCL) language.  However, I would ask Apple to reconsider its decision
to terminate MCL's development in light of MCL's proven and expanding
capability as a revenue generator of hardware sales.  MCL sells MACs: not
for the language itself, but for the applications and their descendants
which are initially prototyped only because MCL exists on the Macintosh.  

There is a Macintosh product a colleague of mine has developed for teaching
radiologists about magnetic resonance imaging.  The product is written in C,
and has directly resulted in the sale of hundreds of Macintosh computers.
It exists for one reason: my group at Carnegie Mellon first built (in two
months) a fully functional MCL prototype which was used extensively in the
field for several years.  Sustained interest in the novel MCL graphical
interface eventually provided the customers and financial backing which
resulted in Macintosh sales.  And, these MACs were generally sold to
Apple-hostile sites, and ultimately spawned further sales of Macintosh
hardware products.  

Many other groups have employed a similar strategy: rapidly prototype in
LISP, and then port the final design to C.  Neuron Data used this approach
in their initial development of Nexpert Object, which resulted in the sales
of tens of thousands of Macintosh units.  With fast processor speeds and
good compilers, though, applications can be kept in the LISP language as
continually adapting programs.  

Our group is currently applying this approach for molecular genetics
Macintosh-based software.  In this niche market, we are developing novel
functionalities in MCL on the Macintosh, which are delivered as MCL
applications.  What may be difficult to appreciate is the extent to which a
dynamic functional language such as LISP truly enables the rapid development
and customization of new solutions, and their rapid translation into working
usable software.  Coupled with execution speed, a phenomenal development
environment, and painless access to the graphical (and other) resources of
the Macintosh computer, MCL is in a class by itself for those developers
concerned with the rapid innovation so vital to continued acceptance and
sales of the Macintosh computer.  Which is why the PowerPC port for
competitive speed of MCL user applications remains vital to the future of
Apple as an innovator.  

The last generation of leaders in software was schooled in C and UNIX, and
helped liberate their shops from the restrictions of COBOL, FORTRAN, and
IBM.  The next generation is now educated with an awareness of the power of
dynamic functional programming languages for dramatically decreasing the
cost, time, and maintenance of customer-driven software applications.
Currently, Apple's MCL is far and away the best such product on any
platform.  It is this competitive edge that will help Apple continue to
expand (over DOS and UNIX) in the burgeoning custom software markets, where
Macintosh-only custom software gets Macintosh computers in the door.  

I therefore urge you to consider only one crucial aspect of the future of
MCL to future of Apple: the bottom line.  Help your Apple developers keep
one step ahead of everyone else.  It is the software which sells your
machines, and with declining profit margins in the software industry, you
need to keep your developers loyal to Apple.  

Thank you for your time.  I would be happy to further discuss with you (or
anyone else at Apple Computer) MCL's importance to the Macintosh product
line.  

- Mark
-------------

Mark Perlin, Ph.D., M.D., Ph.D.
Research Computer Scientist
School of Computer Science
Carnegie Mellon University
Pittsburgh, PA 15213

perlin@cs.cmu.edu
(412) 521-7421
(412) 681-5739 (FAX)