Re: Thou shalt have no other gods before the ANSI C standard

From: rpl (plinnane3REMOVE_at_NOSPAMyahoo.com)
Date: 02/16/05 

Date: Wed, 16 Feb 2005 12:23:37 -0500

jmfbahciv@aol.com wrote:

> In article <qfSdnZ1jkbqVg4_fRVn-2g@rogers.com>,
> rpl <plinnane3REMOVE@NOSPAMyahoo.com> wrote:
>
>> jmfbahciv@aol.com wrote:

>>> I inherited a COBOL program to maintain where the developer
>>> "saved code" by calling subroutines which were paragraphs
>>> in the main body of the flow chart. He assumed that the
>>> fewer ASCII characters in his source would produce "faster"
>>> executables.

>> it's called "Structured Programming"

> This was not structured at all. Let me try to describe what
> was going on further. Essentially the subroutines were PERFORMing
main body paragraphs.

Uppity subroutine. That's not a valid method.

> Now, this concept could be s..t..r..e..t..c..h..e..d to a half-assed
> coroutine concept but that's not how the code flowed.

I can see some near-valid reasoning...

fatal-error handling (should be a GOTO though, PERFORM just buggers up
somebody trying to read the source code)

backtracking (but again, should be a GOTO, unless you want your stack
screwed over and that kind of control belongs to the mainline anyways)

It sounds that the paragraph should have been physically outside of the
mainline and been separately PERFORMed from both the subroutine and from
the mainline. (adds a few bytes here and there, but avoids other issues)

> Again, I'll try to describe how the programmer thought. If the
> _ASCII_ characters of the source was less, then the executable
> instructions generated by the compiler would be less.

technically correct:

PERFORM <section or paragraph> is just

push program pointer
jump to <section or paragraph>
<do the section or paragraph>
pop program pointer
jump(back)

It's not going to be faster no matter what though, and if the PERFORMing
paragraph is sufficiently far away from the PERFORMed paragraph you've
got *major* prehistoric thrashing problems on your hands.

> I just got the feeling that you won't understand what I mean
> by "instructions generated by the compiler".

:(

rpl

Relevant Pages