[Oberon] a module a page (keeps the mind sane)?

[Bob Walkden]

> On 3 Nov 2018, at 14:55, Paul Reed <paulreed at paddedcell.com> wrote:
> 
> Hi Frans-Pieter,
> 
>> Does anyone remember where Niklaus mentions the ' one page of code, one
>> module ' rule?
> 
> When you originally asked this (on this list in December 2015, can't
> believe it was that long ago!) :) I asked a colleague and he couldn't
> remember where he had read it either.
> 
> But asked again and intrigued, he embarked on a browse of his dead-tree
> library and eventually conceded that what he had probably remembered, or
> at least the closest thing he could find, was
> 
> "The Elements of Programming Style", 2nd Edition (Kernighan and Plauger),
> page 150, bottom two lines:-
> 
> "The modularization would have the advantage that each part of the program
> would fit comfortably on one page."
> 

There was a very practical reason for a 'one page, one procedure' rule when I started programming in the early 1980s. 

I was a COBOL programmer and we did not have terminals or such luxuries. Instead we wrote our code in pencil on coding sheets (paper), submitted it to the data prep women (always women) and a few days later received the sheets back with a paper tape version of the source code which we could then attempt to compile. 

Because we practised Jackson Structured Programming we could write subroutines which fit on one coding sheet and re-use the coding sheets for other programs, or shuffle them, insert other code on separate sheets etc as needed. Only for short, simple edits and quick compiler turnaround when we had the machine to ourselves on a Tuesday evening did we need to cut and paste, quite literally, the paper tape. 

Anything involving a new coding sheet would take a week, but that was much easier than dealing with something monolithic which involve re-pencilling numerous sheets from scratch.