Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!wasatch!uplherc!esunix!bambam!bpendlet
From: bpendlet@bambam.UUCP (Bob Pendleton)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Re^2: TrueBASIC (was Re: EZ-DOS, MS-DOS compatible OS)
Message-ID: <128@bambam.UUCP>
Date: 22 Aug 89 16:16:13 GMT
References: <1635@bucket.UUCP>
Organization: Evans & Sutherland Computer Corp., Salt Lake City, Utah
Lines: 66

From article <1635@bucket.UUCP>, by leonard@bucket.UUCP (Leonard Erickson):

> At the time GW-BASIC was written, ANSI Minimal BASIC was ANSI BASIC.

Restate that as "At the time GW-BASIC was written, ANSI Minimal BASIC
was the ONLY ANSI BASIC" and I'll buy it. Back in '79 I was working
from a document that clearly had the words "Minimal BASIC" on the
cover. 

And Yes, I've worked on implementations of both ANSI Minimal BASIC and
ANSI BASIC. Neither product was ever released, but thats the way it
goes sometimes. 

I recieved my first public review copy of the ANSI BASIC standard in
'81 or '82. The ANSI Minimal BASIC standard preceded the ANSI BASIC
standard by several years. 

> It wasn't until a lot later that the current ANSI BASIC standard demoted
> the old one to "ANSI Minimal BASIC".

What can I say, go check the dates on the two standards.

> There are good reasons why micro BASICs's are interpreted. They're
> intended to be used in machines with almost no resources, and *very*
> novice users. 

This is a slam to both PCs and users. A modern PC has more memory and
more processing power than a large mainframe had in the early sixties.
I don't think anyone sells a machine with as little memory as the
machine on which LISP and FORTRAN were originally developed. 

I've found that "very novice" usually means "Not indoctrinated," that
is, unwilling to put up with the garbage we computer folks have come
to accept as normal. 

> Can you picture trying to have the typical *home* user
> do an edit/link/compile cycle? 

Nope, I can't. Do you really believes that the only way to use a
compiler is to go through an edit/link/compile cycle? Have you ever
heard of incremental compilation? How about fast memory to memory
compilers? How about parse on entry with demand code generation. How
about dynamic linking? They all work. They all give you compiled code
execution speeds without an edit/link/compile cycle. 

> And the PC started outas a *cassette*
> machine... don't even think about compiling in *that* environment!

My first personal computer was a 2MHz Z80 with 32K of RAM and a 1200
baud cassete interface. It had Microsoft 8k BASIC in a plug in ROM. I
wrote 2 compilers and a LISP interpretor in BASIC for it. Even the
memory to memory compiler was pretty slow. All this just to say it
isn't unthinkable and wasn't unthinkable back in those bad old days. 

When was the last time you saw a PC with a cassette? An 8088 with 256K
and 1 360K floppy is all you really need to run an interactive
compiler. Assuming the compiler is designed and written to conserve
resources. 


			Bob P.
-- 
              Bob Pendleton, speaking only for myself.
UUCP Address:  decwrl!esunix!bpendlet or utah-cs!esunix!bpendlet

           Reality is stanger than most people can imagine