Re: Kyan Pascal - new insights

[zcaa1122@rpool9.rus.uni-stuttgart.de (Wulf Hofbauer)]

In article <1993Jul20.200732.43028@urz.unibas.ch> hochstrasse@urz.unibas.ch writes:
>The memory limitation is a general problem on 64K Apple IIs programmed in
>high level, compiled languages. The use of the 2nd 64K bank helps somewhat,
>but to do serious programming in Pascal [I mean applications that really
>do something useful] you need at least a IIGS. It's also more fun, because
>you can concentrate on your project without worrying about memory...
>I think, Kyan offered also a package to use the features of 128K Apple IIs,
>i.e. more memory, Double Hires etc.

Well, I'm doing all of my HLL programming in HyperC. The interpretet cp-
code (similar in concept to the UCSD p-code) is ultracompact and sur-
prisingly fast. And native code is still MUCH more efficient than the
code produced by Kyan Pascal. Btw, the HyperC compiler is written in HyperC
itself. I doubt you could do this with Kyan Pascal...

As to optimization: the Kyan compiler produced a macro for every standard
function called which I consider clumsy. The code was pure stack-machine
code - without ANY optimization. I guess the compiler uses syntax-driven
code generation. On contrast, I tried to write a routine in the
assembly language of HyperC's cp-engine. Though the C compiler doesn't
optimize perfectly, it enden up with smaller code that I produced manually.

I don't want to flame Kyan Pascal. It's just I was VERY disappointed with
it.

- Wulf