Re2: MCL Framework & Direction

[Vincent Keunen <scosysv!keunen@nrbmi2.ia.nrb.be>]

    Date: Thu, 25 Jun 1992 18:25+0200
    From: D1104@AppleLink.Apple.COM (Computer SW Consult, J Stulin,PAS)

    Luke,
 
    You are correct; we will always want more memory, disk, and cpu power, and C++
    will always be more lean and mean than Lisp. But consider this:
 
    1) The price of software development continues to grow, the cost of hardware
    resources contiunues to collapse. MCL development is orders of magnitude
    cheaper than C++ development and always will be, even when the next generation
    of C++ development systems become available.

Seconded.  What needs to be considered is not "pure" speed or "pure"
resource needs.  Otherwise, why doesn't everybody program in assembler:
it's faster than anything else and requires a lot less resources... It's
a compromise between costs: costs of HW versus costs of SW development.

In my view, the language you use directly impacts the things you express
in that language.  Before the mathematical language, there was no math:
how can you easily express calculations, theorems,... in natural
languages?  Of course, you *can*, but because it's not easy, people just
don't do it (we live in a real world with real people - well, some are
more real than others maybe ?! :-\).  Lisp and other languages allow you
to express different things.  It's not fundamentally different, because
you can do everything in binary, at the end.  Everything gets compiled
to the same machine code, so it can't be that different, right?.  But on
a practical point of view (and that's what matters), some languages are
more apt to express certain things.  That's where I see lisp differ from
other languages.
 
    3) Dylan, and a good application class library for Dylan, are vaporware. MCL is
    here today.

Right.  And the technologies don't live by themselves.  They live in a
market.  And the market evolution is orders of magnitudes slower than
that of technologies.  The mac will still be here in 10 years (see the
wonderfull technological evolution and there is still 60 millions DOS
PCs out there and worse - people use them).
 
    Jeffrey Stulin
 

vk

********************************
Keunen Vincent	       	       	       	       	  
R&D, Software Engineer				  

keunen@nrb.be 	

tel: +32 41 407282
fax: +32 41 481170

Network Research Belgium
Parc Industriel des Hauts-Sarts
2e Avenue, 65
B-4040 Herstal
BELGIUM
********************************