From: Dave Thomas <Dave@...>
Date: 2000-02-14T21:08:06-06:00
Subject: [ruby-talk:01360] Re: Say hi (bis)


matz already answered some. I'll take a crack at some others.

Pixel <pixel_@mandrakesoft.com> writes:

> --------------------------------------------------------------------------------
> bad points:
> 
> -- redefine a function -> no error

It used to with -w, but it was a pain (because many methods are
intended to be overloaded). However, it's easy to produce a mix-in
module that checks for redefinitions whenever a method is added to a
class. This could be applied globally or to a particular class.

> -- no type checking (really, there could be a mode paranoid that
>    would catch most typos)

Personally I think that's a big plus ;-) However your could write (in
Ruby) a set of functions that (say) enforced types on a collection
class: given a template class you could create new classes with the
template instantiated.

> - hash are real hash -> can't keep order (maybe a Vector would be
>   nice (also more economic))

There's also the assocc/rassoc functions.

> - doc not as good as perl's :ppp

We've recently started working on a new Ruby book, and this list is
helping tighten up the English-language FAQ. However, you _are_
something of a pioneer using Ruby right now. That has its good points
and its bad points.

> / in hash, { a => 1 } not allowed, must be { 'a' => 1 }

One of the joys of Ruby is that it's far, far more regular than
Perl. When you see 'a', you know it's a local variable or a method
call. I must admit the 'a' => 1 thing bugged me a tad at first, but
I've come to appreciate the tidiness of it.

> / no implicit transformation string <=> num

I could argue this one both ways, having been bitten by Perl doing it
and Ruby not doing it. In the end, I think I come down on Ruby's
side--again it's a question of being regular.


Welcome to the list--and thank you for the thoughtful points.

Regards


Dave