Elias, Juergen
I use the "Toronto Toolkit" convention, which looks like this:
∇y←x adjust d;⎕IO;ex;i;line;lmrg;pw;w
⍝adjust each row of matrix <d> according to parameters <x>
⍝.e ('/' ∆box 'please do not / enter') = 15 adjust 'please do not
enter'
⍝.k formatting
⍝.n rml
⍝.t 1992.4.24.14.4.17
⍝.v 1.0 / 05jan82
⍝.v 2.0 / 05apr88 / change order of <x>, use subroutines
⍝.v 2.1 / 24apr92 / using signalerror
⍝ x[1] width of result in columns
⍝ x[2] width of left margin (i.e. number of blank columns)
⍝ x[3] number of blank lines to insert between each row
⎕IO←1
...and more of the function...
I suspect that if the first line is a lamp line, that can be taken as
the description, and a double-lamp taken as a single lamp. This lamp
line may be indented, and could then be the "help" for the function.
The "Toronto Toolkit" format then has lamp lines with ".x" (. followed
by letter) and data .e is example, .t is timestamp, .v is version, .k
for keywords, .n for name/initials and more. Then there are function in
the Toolkit that handle this data. But, again, the first lamp-line gives
the overview.
Just food for thought.
FredW
