On Mon, Sep 25, 2000 at 03:30:47PM +0200, Bart Lateur wrote:
> If you can garantee that it's also not using a hash internally to keep
> count, but instead a table parallel to the table that's being used to
> hold the conversion values, you've won me over.
Naturally, it's hard to guarantee anything since this isn't implemented yet,
tr/// hasn't happened yet, and not a single line of Perl 6 code has been
written. It's a SMOP; "use an array instead of a hash". This changes nothing
about the design.
It seems I was wrong - when I did implement this before, I did use a hash:
http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2000-02/msg01192.html
but a parallel table could have been possible and very easy.
Incidentally, so what if a hash is slow? You pay for what you get. It's still
quicker than doing it by hand.
--
"I will make no bargains with terrorist hardware."
-- Peter da Silva
