Re: "The Best Programming Language for the End of the World"

In article <7ab907de406b3b22270e4118b727c265@www.novabbs.com>,
mhx <mhx@iae.nl> wrote:
<SNIP>

Unfortunately (?) my approach only works for small, dedicated
programs. There are examples where I wanted to write a bigger
application (e.g. MANX, SPIFSIM, SYSSIM, iSPICE, ..). I am
constantly trying to refactor those programs in small stand-alone
modules but I have failed (e.g. MANX which has horrible
object-oriented rubbish) when I (perhaps temporarily) lost
interest in the subject itself.

I felt bad when I ditched MANX and gave up maintaining it.
Following your advice I have rewritten it, and it is now
healthy with a much simpler object oriented approach
(one screen, but it pulls in the FORMAT&EVAL, also one screen).
More powerful too, happily playing sustained notes (organ, pc-speakers)
percussion (metallophone, drum) and a midi expander, at the same time.

I have done much maintenance, and had succes with slash and burn
maintenance, i.e. removing bad features before adding a replacement.
Also I documented features and through the attempted specs I discovered
tests that failed, that in turn forced permission to change code.
manx and eldo (Dutch taxes documentation system) were the exceptions
where this was less succesful.

>
As I am also a ngspice maintainer (19,161 'C' Files in 4,856
directories, only user documentation), I know my Forth
approach works (iSPICE has only 7 dedicated files).

These kind of projects are a nightmare. For example after two
years of eldo I only succeeded in partly documenting how it worked.
The "documentation" was 1 meter of binders, virtually worthless.

>
-marcel

Groetjes Albert
--
Temu exploits Christians: (Disclaimer, only 10 apostles)
Last Supper Acrylic Suncatcher - 15Cm Round Stained Glass- Style Wall
Art For Home, Office And Garden Decor - Perfect For Windows, Bars,
And Gifts For Friends Family And Colleagues.