[erlang-questions] Parallel Shootout & a style question

[Isaac Gouy]

--- Kevin Scaldeferri <kevin@REDACTED> wrote:

> 
> On Sep 2, 2008, at 12:14 PM, Isaac Gouy wrote:
> 
> > On Tue, Sep 2, 2008 at 6:19 PM, Kevin Scaldeferri wrote:
> >
> >> The problem spec doesn't actually require a specific output, or
> >> ordering of that output.
> >
> > Every problem spec begins with an explicit demand that you diff  
> > program
> > output with a provided expected output file - which does seem
> awfully
> > like requiring a specific output.
> 
> Kinda, although the wording makes it sounds more like "don't submit  
> broken code" than "you must output this exactly" to me.  Anyway, my  
> latest version keeps the same ordering, so it's a moot point.

How can diff tell us that there is broken code if it isn't the case
that "you must output this exactly"?


> >> But, the shootout maintainers have a history
> >> of changing the rules retroactively, so it does seem like a good
> idea
> >
> >> to try to keep the same order.
> >
> > The maintainers have a proud history of changing the rules
> > retroactively - some people like to characterise that kind of  
> > behaviour
> > as learning from your mistakes.
> 
> And some people have other opinions ;-)

People tend to have opinions, some justifiable some less so.

The alternatives seem to be:
- never make a mistake
- make mistakes but don't fix them
- make mistakes and fix them.

Would you really prefer that we simply didn't fix the mistakes we make
with the rules?