[ruby-talk:02018] Re: no Fixnum#new ?

From: Yukihiro Matsumoto <[e-mail]>
...
>     on 00/03/17, "Conrad Schneiker" <[e-mail]> writes:
>
> |> I like init too. If we do go with it though, we'll need something like
> |> a full release cycle where Ruby warns if it comes across an instance
> |> method called init--something like:
> |>
> |>   warning: in the next release of Ruby method 'init' will acquire
> |>            mystic significance and this code probably won't work
> |>
> |> Just to give people time to switch over.
> |
> |I also want to vote for 'init'.
>
> Hmm, thank you for votes.
>
> But changing a fundamental method name like `initialize' makes me
> nervous.  It may require thousands of programs to be rewritten.

Well, you may want to save these sorts of things until such time that it
seems that you have accumulated enough of the "I wish I had done this
originally" sorts of things that you think that most people would accept a
Perl4 --> Perl5 or Python 1.6 --> Python 3000 sort of transition. Although
it was certainly temporarily inconvenient for many people, the combined
simplifications and extensions of the Perl4 to Perl5 transition ultimately
boosted Perl's popularity and utility overall. (In addition to Ruby syntax
changes, other things that might also be changed at such a major transition
for Ruby might be a switch from K&R C to ANSI C/ISO C++ and/or the addition
of Unicode 3 support, and so on.) If we knew that (and what) various
language changes would be forthcoming a year or two from now, there would be
plenty of time to give some thought to developing tools that would
facilitate upgrading old code, and to document such things well in advance.

If Ruby undergoes the same sort of long term growth as Perl did, then the
vast majority of Python code that will exist 10 years from now has not yet
been written, so it is important to keep in mind the importance of language
improvements that would make the development and support of such code easier
and that would improve the overall quality of that code. If Perl had been
forced to forever remain perfectly backwards compatible with Perl4, Perl
would not be nearly so useful, powerful, and easy to use (at least for those
that like it) as it is today.

Conrad