Symbolics' attitude problem

[Chris Lindblad <cjl@WHEATIES.AI.MIT.EDU>]

    Date: Fri, 22 Jul 88 23:57:47 PDT
    From: gyro@kestrel.arpa (Scott B. Layson)

       Date: 22 Jul 1988 01:15-EDT
       From: NCRAMER@G.BBN.COM

	  > Date: Tue, 12 Jul 88 10:52 EDT
	  > From: Scott McKay <SWM@SAPSUCKER.SCRC.Symbolics.COM>
	  > Subject: Re: COMPILE-SYSTEM, AGAIN
	  > I realize that this sounds arrogant, but just because some users (not
	  > "most users", as you claim) "want" a certain behavior is not sufficient
	  > justification for doing it.

       To the more general (and the real) question at hand: what can I say?
       Yes, it does sound arrogant and more than a little so.  It would be
       interesting to know what _does_ constitute "sufficient justification"
       for the inclusion of a feature in a product when the desire for it by
       the knowledgable users and customers of the system does not.

    It doesn't just *sound* arrogant -- it *is* arrogant.  If Symbolics is
    ever to survive it is going to have to get over its "we know better
    than the customer" attitude.  That and its narcissism are what has
    gotten it into its present straits, and will likely lead to its
    demise.  And quite a shame it will be, too.  If laying off half your
    workforce hasn't taught you people some humility, what will?

Now hold on.  I have found that Symbolics is the most responsive computer
company I have ever dealt with.  With what other computer company can you
actually discuss in a meaningful way over electronic mail the direction of
development at the company?  With other companies I don't even bother trying
to discussing product features with the developers because I know it will get
nowhere.

How many UNIX computer companies are willing to make improvements to the Make
program?  The fastest growing computer company today is also the most
arrogant.  When was the last time Bill Joy actually asked anyone what he
wanted.  Here's a quote from an interview of Bill Joy with The Sun Times, "A
news and information journal devoted to the Sun Microsystems Community"
(emphasis is mine):

    "I think Unix and OS/2 are probably going to be the two dominant operating
    systems in the 90s.  Where OS/2 is the ting that runs on all the IBM PS/2
    clones or whatever, and Unix runs on everything else.  1There will at some
0    1point probably be a new system, a new avant-garde sort of operating0 1system,
0    1which no one will take seriously but me; eventually it will be important
0    1just0 1like Unix.0"

And when was the last time anyone ever got a copy of the Lucid sources?

The reason Symbolics is having a hard time probably has less to do with
responsiveness to customers and more to do with:

1.  Lisp Machines don't run UNIX.
2.  The cheapest Lisp Machine costs more than the cheapest box that runs UNIX. 

Scott has said that he has come up with a solution to the recompiling system
problem that is different to the solution that was suggested.  That's good
enough for me.