From: Sean Middleditch <elanthis@...>
Date: 2002-04-05T04:01:24+09:00
Subject: Re: About efficiency

On Thu, 2002-04-04 at 12:00, Yukihiro Matsumoto wrote:
> Hi,
> 
> In message "Re: About efficiency"
>     on 02/04/05, Sean Middleditch <elanthis@awesomeplay.com> writes:
> 
> |> In ruby, identifiers are handled as numeric ID commonly.  Don't
> |> you mean this?
> |
> |Sort of - are the names converted to ID's at compile time or run-time?
> 
> Compile time.

Ah.  Is it guaranteed that all names which are the same will have the
same ID, then?  That would speed up run time, but could be a pretty big
slowdown on compile time for large apps... (I think, guess I'd have to
try it both ways. ~,^)

> 
> |Well, with a cache, you wouldn't need to know the order until first
> |encountered.  If you don't allow the removal or deletion of members,
> |then once cached, the index will always be valid.
> 
> Unfortunately, Ruby allows method removal.

Hmm.  Well, that would make it a tad bit more difficult.  ^,^

> 
> 							matz.