[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

the future of MCL

We feel that we must add our 2 cents worth concerning the future of MCL
(in the hope that the mere quantity of responses will get someone's attention).

Since it was first marketed by Coral, we have been using MCL to explore the use
of Lisp as a tool for time series analysis.  We started out in the mid-80s
working on Symbolics Lisp Machines, but switched to Coral's product when
it became available because the price of the software + the necessary
hardware was MUCH less than a one year maintenance contract for the Lisp
machines.  The "necessary hardware" consisted of two Mac II's in 1988,
which have since been replaced by a Centris and a Quadra, so Apple can chalk
up about $15K in purchases that were justified SOLELY by the existence of MCL
(OK, no big deal -- BUT, if you consider the number of responses to
the "death of MCL" posting with stories similar to ours, we're starting
to talk about REAL money).  However, the story doesn't end there:

[1] Because these Mac II's were the very first in our organization,
    they generated a lot of interest, indirectly influencing the purchase
    of many more machines (our organization now has well over 100 Mac's).
[2] Several organizations purchased one or more Mac's + the MCL software
    to run our time series analysis system (including the Naval Research
    Laboratory, the Naval Observatory and the [at the time] National Bureau
    of Standards), so chalk up at least 5 more machines purchased
    solely due to MCL.  For some of these organizations, these were
    the VERY first purchases of Apple products, leading undoubtedly
    to other purchases that we are not aware of.
[3] We have used our time series analysis software in a number of graduate
    level classes over the past five years, introducing a fair number of
    grad students to MCL and to the Mac.

Finally, we would like to note that Apple products and MCL are being
considered for a key role in upcoming projects we are involved in, including
two (perhaps three) multimedia training products.  We have reviewed the
current crop of training system authoring tools and found them to be
lacking in scripting (programming) and data structures (objects).
Since we want our systems to be significantly more capable than
the simple training systems to which most people have been exposed (we
want to seamlessly integrate a variety of training strategies
and background reference material), we need a more powerful tool kit.
When we boil our authoring requirements down to their essence, we find
that we need good data structures, lots of object relationships,
access to outside resources (e.g.  databases, connectivity, media, ...),
good screen presentations, and a friendly user interface.

We are considering Common Lisp + CLIM because they provide most of the needed
capabilities now.  They are not a hollow promise of capability to be
delivered "next fall" (although we are very disappointed in the way the
CLIM implementation for the Mac has been handled so far, and we are now
worried that MCL might disappear).  The language we select must be fully
supported and enhanced, with a growth path to new platforms.

Lisp/CLOS is a true object-oriented language (not an approximation++).
This helps make Lisp the most extensive, extensible, and productive
language available (see the preface by Ike Nassi in the Dylan manual).
Apple's hardware + the MCL software is finally delivering the performance
promised by Lisp on affordable platforms, yet inexplicably Apple is now
abandoning MCL.  If Apple believes the preface to its Dylan book,
then it should support MCL and CLIM and not commit to a pie in the sky.

And now for the cost/benefits.  If these training products are successful,
we expect to produce thousands of units per year.  Apple's commitment to MCL
will determine if those units will run on Apple products ... or not.

We close by repeating what Rainer Joswig said (and what should be the
bottom line from Apple's point of view):
                      MCL sells expensive Macs   ...
                      MCL sells expensive Macs   ...
                      MCL sells expensive Macs   ...


Don Percival, Greg Anderson and Keith Kerr
Applied Physics Laboratory
University of Washington
Seattle, WA

PS:  In view of Bob Futrelle's signoff ("Sleepless in Edinburgh"),
     we would add "Sleepless in Seattle", but strangely someone
     seems to have a copyright on that slogan ... :)