PowqerPC port: can we help ?

[Marc Domenig <domenig@urz.unibas.ch>]

------------------------------ Start of forwarded message 1

Delivery-Date: Sun, 20 Feb 1994 09:00:30 +0100
X400-Content-Type: P2-1984 (2)
X400-Originator: owner-info-mcl@cambridge.apple.com
X400-Recipients: domenig@urz.unibas.ch
Date: Sat, 19 Feb 1994 22:03:10 +0100
From: twl@cs.brown.edu
Message-ID: <9402200703.AA10433@lute.cs.brown.edu>
To: Empty Address:;, nassi@cambridge.apple.com, info-mcl@cambridge.apple.com
Cc: twl@cs.brown.edu
Subject: Future of MCL
Content-Length: 4977

Dear Mr. Fleischman,

I was very disturbed by your message regarding the future of MCL.  I
am aware that Apple is under financial pressure and that the company
want products to pay for themselves.  The MCL situation is an specific
instance of a problem that seems to be growing at Apple.  Developer
tools should not be a profit center.  I am amazed that the management 
at Apple cannot see that.  I am even more amazed that this attitude
can persist with the transition to PowerPC less than a month away.
Apple needs to attract innovative developers to the Macintosh
platform, now more than ever.  Application differentiation between the
Macintosh and Windows is shrinking every day.  I understand that Apple
has responded to this need by creating an in-house group that is
charged with developing applications that showcase the benefits of
Macintosh system software. Why then does Apple seem to be reluctant to
invest the resources needed to produce developer tools that will
attract developers, or at this point, keep them in business?  At the
moment, there are no viable C++ compilers for the Macintosh.  There
are two versions of the MPW C++ compiler, which are slightly
incompatible.  The highly publicized "partnership" with Symantec has
resulted in C++ compiler with a huge public bug list, and a vaporware
cross-platform framework (Bedrock).  The Metrowerks C++ compiler may
be the only hope for a reliable and stable C++ on the Macintosh, but
it is still unfinished.  Meanwhile vendors of Windows developer tools
are pulling out the stops to produce better and better environments.

Now we are told that MCL will not be ported to PowerPC, and that
features will not be added because the sales of MCL do not cover the
resources require to do the work.  MCL is a development platform that
supports rapid development of highly innovative applications.  It is a
common practice in the academic and industrial research community to
build research prototypes using Lisp, and then deliver production
versions in some other language.  MCL also provides an easy mechanism
to incorporate AI research into Macintosh applications.  This alone
should make MCL a strategic product for Apple.

MCL attracts developers to the Macintosh that will not be attracted by
any other Macintosh development tool product.  MCL is the *best*
Common Lisp implementation on stock hardware.  We have applications at
Brown that crash Lucid Common Lisp, but run fine under MCL.  If
marketed properly, MCL on PowerPC would provide a very cost-effective
alternative to running Commmon Lisp on a UNIX workstation.  MCL on PPC
would cause the market share of MCL overall to increase dramatically.
There is no product for DOS/Windows that compares to MCL.  This is an
advantage for Apple, and should not be thrown away.  There is no
Common Lisp running on stock hardware that can beat MCL.  I urge you
to consider a revised marketing scheme.  Perhaps luminaries like Henry
Lieberman would be willing to give you a product endorsement. There is
enough of a market for Common Lisp to support Lucid, Franz, and
Harlequin.  There's plenty of room for a product of MCL's caliber.

Perhaps Apple intends for us all to switch to Dylan.  Unfortunately,
the language design is not finalized.  No one outside Apple has
written a single shipping application in Dylan. Not one word has been
said about any implementation of Dylan being available from any
vendor.  The Dylan development team is much bigger than the team
working on MCL.  How is this product paying for itself?  Dylan is not
a sure thing.  All the copies of the Dylan book that have been mailed
do not translate into sales of Dylan implementations. Realistically, it
will be years before Dylan has the kind of acceptance that Common Lisp
has.  Realistically, it will be years before the first Dylan
implementations are as mature as MCL 2.0.  I am not bashing Dylan --
part of the reason I want MCL is to have an environment to prototype
Dylan variants with.   Even if Dylan gains wide acceptance, there will
still be a place for MCL.

I have been delaying my purchase of a personal copy of MCL in order
to find out about the future of MCL.  If MCL development does not
continue, Apple will lose a sale -- either they will lose the sale of
a PowerPC upgrade, or the sale of a copy of MCL.  And perhaps people
like me will be forced by Apple to switch to another platform. I hope
that none of these things come to pass.  I'll put my money where my
mouth is.  I guarantee you one additional sale for the PowerPC version
of MCL 3.0.

I hope that Apple management can be persuaded to reconsider their
decision regarding MCL, or that a suitable development partner can be
found.  I also hope that such a partner will prove to be more reliable
than Symantec has been.  Perhaps the MCL situation can be resolved in
a way that inspires confidence in the Macintosh developer community
instead of fear. 

Ted Leung
Brown University Object-Oriented Database Group

------------------------------ End of forwarded message 1