Re: New style DO syntax?

Greg Lindahl <lindahl@xxxxxxx> wrote:

Language additions have a higher hurdle than you think, especially when
they don't add anything but your idea of better aesthetics.

Yes. That's partly why I've been mostly staying out of this. I can't see
much chance for most of these proposals. It is hard to get agreement
over what is or is not an aesthetic improvement. But even assuming there
was widespread agreement on it, one tends to need more justification
than that.

After all, there is zero - no, make that less than zero - chance that
the existing DO syntaxes will be removed. We already have two different
syntaxes for even basic looping (statement label and enddo forms), not
counting the variations like while. If you add another, that makes the
language more complex, because the old forms will still be there. So
there will be more forms to specify in the standard, more forms for
compilers to handle, and more forms for a reader to be familliar with
and wonder what the differences mean.

So any possible benefit in terms of lower error rates (that's one of the
things often cited for some things), is likely to be counterbalanced by
higher error frates because of people confusing the multiple forms. Not
to speak of the issues relating to having the standard specify how all
the forms interact when nested.

I'm not going to get into whether I do or do not think any of the
proposed forms are improvements. I haven't actually spent much time
thinking about it. I just haven't heard anything that sounds like enough
justification to fly.

--
Richard Maine | Good judgement comes from experience;
email: last name at domain . net | experience comes from bad judgement.
domain: summertriangle | -- Mark Twain
.