Re: Creating and using arrays - again

Andrew Haley <andrew29@[EMAIL PROTECTED]
> wrote:
> Jonah Thomas <jethomas5@[EMAIL PROTECTED]
> wrote:
2...
> > If you're willing, look at how you'd get nearly the same effect
> > without
> > DOES> . There are a few Forths that don't give you DOES> and someday
> > DOES> you might use one and need to work around it.
> 
> Wouldn't it be better to write DOES> for such a Forth?
> 
> DOES> has been in Forth for a very long time.  A Forth without DOES>
> is -- IMO -- fit only for the round filing cabinet in the corner.

First off, if you aren't real familiar with Forth, then inventing
something like the approach Mark Humphries provided will show a lot
about what DOES> accomplishes. You can get by without it with a little
extra effort and using a little extra memory. DOES> is a nice
convenience but it isn't a necessity.

Second, there are trade-offs to implement DOES> in high level code. The
best way I see involves an extra inefficiency in every CREATEd word in
exchange for an easy DOES> on the ones you use DOES> with. It's worth
knowing about the subtleties involved even if you don't ever have to
deal with them.

Third, a close study of DOES> will leave you dissatisfied with it. You
could easily have an approach that's far more flexible and perhaps more
useful, and even easier to implement, except that we have the
traditional DOES> as a tradition. That's worth noticing too.

So I say it's worthwhile to look at it, even if you then always use
DOES> instead of any alternative.