🚀 go-pugleaf

In article <3079@sequent.cs.qmw.ac.uk>, mmh@cs.qmw.ac.uk (Matthew
Huntbach) writes:
|> >
|> Suppose the point which causes the program to crash occurs
|> somewhere where it is very difficult to track down, and will
|> require months of programming to solve. A version of the
|> program released for experimental use might result in the users
|> deciding that the bit which causes the crash isn't important
|> and can be left out. The implementors are then freed from
|> wasting their time debugging something that wasn't useful
|> anyway.
|> 
|> Releasing experimental languages before the code is fully
|> debugged is like getting your papers refereed. It's better for
|> the research community that interesting papers are debugged
|> through the refereeing process than that research results don't
|> get heard about for years because the researchers are spending
|> all their time trying to get their papers to the point where
|> they can be published without alteration.
|> 

I heartily disagree!  If you've got points in your program that require
months of debugging and modifications to fix, then you haven't written
the program: you only kidded yourself that you had! And if you need that
kind of experiment to find out what is important, then you have no
business implementing the language: design it first!

The refereeing process can often improve the paper, but as a referee I
HATE to have my time wasted by people who just don't care about how they
write. Write it as well as you can, then submit: don't worry, there will
be enough comments anyway!

A badly written program or paper contributes very little and wastes
other people's valuable time. If one doesn't strive for excellence,
one's activity is probably more harmful than beneficial.

-- Feliks Kluzniak