Re: Where to buy compiler

[smm@bix.com (smm)]

David Weller (dweller@Starbase.NeoSoft.COM) wrote:

: Getting shelf space for an Ada compiler is next to impossible.  Hell,
: even finding Smalltalk on the shelf is difficult.  Alas, there's even

: (or even a denial).  In any case, the chance of a "low cost" compiler
: for the PCs does't look very promising (All you vendors: Don't follow
: up to this post and say "We think $999 is low cost for a PC" :-).

That's bad news. I'm glad you mentioned Smalltalk. There's a "non-C"
language where at least one vendor _DOESN'T_ _HAVE_ _A_ _UNIX_
_PRODUCT_ (Digitalk) and they've been in business and heavily hacking
and improving their product for quite some time. I could be mistaken,
but I don't think IBM's new Smalltalk supports Unix either (yes, NEW)
but at least all their advertising talks about Windows.

I don't know whether to laugh or cry.  Folks peddle brain damaged
ports of language Z to the PC that typically (1) cost way too 
much and (2) require oodles of RAM, often to compile, much less
write a program, (3) charge license fees for distributing
programs or (4) are small and crippled for "educational use". And
the peddling isn't very good. No add's in popular trade rags, PR
targeted at reviewers, columnists, etc.

Meanwhile a $89 C++ compiler supports Windows, has a giant
class library devoted to making Windows simple, has a visual
programming environment, has a GUI builder, has a slick 
debugger, AND (for techies) has what many call is the slickest
templates implementation on the market.

And the "non-C" language community has the balls to boo hoo hoo
over why their language is "obscure".  Well no shit. WAKE UP.
The mainframe is dying. The mini is dying. Monolithic frame
works are coming and "portability" is going or will be achieved
only at _enourmous_ cost (like Taligent's efforts).  The language
becomes increasingly a vehicle to access pre-canned services
(DBMS's, mail API's, etc.) It's all "bindings" and it will be
a scramble just to _KEEP UP_ with the API explosion on the
popular platforms.  OpenGL comes to Windows and you know where
the slickest supporting tools/OOP Wrappers/etc. will be in
a years time. You've got teams devoted to this stuff at MS,
Borland, IBM. (Borland does this kind of work for Pascal, a
"non-C" language).

I happen to be a BIG NON-C LANGUAGE ADVOCATE. But with this
sort of vendor behavior, we might as well GIVE UP. Where the
hell is the "OLE-2 Component Design Kit" based on Eiffel,
ADA, Common Lisp, Scheme, ML, Sather, Modula-3.  This is
the _SWEET SPOT_ of the market place. This is the stuff
a bunch of ex-accountants will learn so that they can go
train MIS shops. This is the stuff that makes it to Crown
Books book shops.  This is the stuff that crosses the desk
of managers in their trade rags.

Meanwhile, a bunch of us nerds pat ourselves on the back
because the nifty language Z now compiles on the Alpha
CPU, not just the Sun (but not Solaris). The debugger
will be done in 6 months.  The MIPS port lost funding. Ha
ha ha.

You want to stroke desk officers? You want to run on 1K little
chips? You want to teach how to write a better "hello, world" 
to college freshman? You want to play around with better 
register allocators?  You want to make sure the nuclear
reactor doesn't blow up? Ok. Then use ADA, Forth, Scheme,
Modual3 or Eiffel.  This is what the market is telling us.

You want to make a company that sells an E-Mail package to
pre-teens with Internet accounts? You want to do contract
programming for the giant market of Windows users? You 
want to write a DLL that ferrets the database connection
handle out of the client runtime system because the debug
messages are faulty (a real recent problem, but typical of
C/Pascal hacking in MIS shops) -- then you use C/C++/Pascal
-- That's what the market is telling us.

IT'S TOO BAD BECAUSE FOR MOST TASKS C/C++/PASCAL SUCKS! 

This just DRIVES ME CRAZY!  The vendors need a big kick
in the ass and a giant dose of Wall Street and a giant
dose of Main Street.