Symbolics Germany prices

[Reti@STONY-BROOK.SCRC.Symbolics.COM (Kalman Reti)]

I'm making these comments as a private person who like the Symbolics
environment, not as any sort of spokesperson for Symbolics (although
I do work for them).

    Date: Tue, 27 Aug 1991 10:17 EDT
    From: %nrb.be%mcsun.EU.net@Warbucks.AI.SRI.COM

    Date: Tue, 27 Aug 91 16:16+0100
    From: Vincent Keunen <keunen@milou.nrb.be>
    Subject: Symbolics Germany prices
    To: SLUG@WARBUCKS.AI.SRI.COM
    cc: uucp%"info-mcl@cambridge.apple.com"@nrbv01, eric@milou.nrb.be
    Message-ID: <19910827151626.1.KEUNEN@milou.nrb.be>
          
	      Date: Mon, 26 Aug 91 13:37 CDT
	      From: Shepard@MCKENZIE.S&C.Dialnet.Symbolics.COM (Tom Shepard)
          
          
	      It all depends on how long you intend to be programming, or how long
	      management intends to stay in business.  If management only makes
	      decisions based on absolute cost, maybe they don't know how to make
	      Return-On-Investment calculations.  
          
    Our company is doing well, thank you (it's not because of the symbolics
    we have, though).  Still, our management (which is also taking the
    decisions for the other parts of the company that bring money, ie
    IBM9000 & so on) doesn't fully agree/understand your simple formulas...
 
    I agree the Genera environment is very good and I like it very much.
    However, it's still far from ideal, when I consider the *weeks* we
    spent debugging dna and recently cterm - just because Symbolics Germany
    doesn't even know what these packages are for...  But they know what
    software maintenance for dna is (in terms of DM I mean)...
Yes, unfortunately, in the past it looked like there would be resources to
develop and support more products than turned out to be true; many people
within Symbolics have argued that products like DNA should simply be
dropped completely because they aren't even close to finished.  If Symbolics
Germany gave you the impression these were polished products, someone was
mis-informed.  (I'm not pointing a finger, it may well have been that
Symbolics Germany didn't really know this.)

There are many products which are not particularly well supported by other
vendors and even dropped.  (For example, my wife has an application with
require MacIntalk and which therefore prevents her from upgrading to System 7).
It is unfortunate you ran into one of those on the Symbolics machine, but
that isn't an argument for knocking the useful parts of the Symbolics environment.

(If you had gotten a similarly unfinished DNA/CTERM product from another vendor,
you probably wouldn't have had the opportunity to spend the time to make it work.)

    If Genera is better AND much more expensive; what's the big deal?
    Having a better environment for the *same* price - that would be much
    more appealing!  And wouldn't it allow Symbolics to get a critical
    market share, which it needs badly for lowering production costs?
Unfortunately, one of the reasons for the extra price is the support.
It costs more to support a large system than to support a small system.
Notice, you cannot buy anything comparable to Genera, i.e. an entire
operating system written in Lisp for which 85% of the sources are shipped
and user-modifiable.  In many of the projects I've worked on, the ability
to modify behavior normally reserved to the operating system (and therefore
normally completely out of the reach of the application) has been crucial
to success.

    Now, in your formulas, just put these elements:
 
    - MacIvory + Genera = $38000
       (plus Macintosh price)
But you are also getting a wholly separate computer, with virutal memory,
disk, network, scheduler, utilities, etc.  That is hardly comparable
to adding Common Lisp to your existing computer!!
 
    - Macintosh Common Lisp = $400
       (plus Macintosh price)
 
    Now do you really think Genera + Ivory is worth 95 times MCL???
 
    MCL is definitely not as good as Genera, but it still is cltl2
    compatible (although beta), 
Does it handle a gigabyte virtual address space?  If so, when you GC
that address space, what does it do to performance?  Does it support
object oriented programming efficiently? 
				has good programming utilities (thru M. Kant
    work, notably, like defsystem, who-calls database, a metering system, a
    compare sources,...), a much better interface designer than Frame-up,
    access to Macintosh resources like the communication Toolbox for easy
    networking, a pretty good emacs-like editor, a good compiler (soon
    incremental), etc...

    Now do you *really* think Genera + Ivory is worth *95* times MCL???
Of course, it is obvious that it depends upon what you are doing.

If you are building toy systems or learning how to use Common Lisp, clearly
not.

If you are building production transaction-oriented applications, where
expert systems, heterogenous network communication, database access and
a lot of traditional data-processing-like code need to 1) be developed
quickly, 2) be thoroughly integrated and 3) need to track a complex and
and/or quickly changing external environment, you bet your socks!

    Vincent Keunen
    keunen@nrb.be