From: Christian Szegedy Date: 2001-05-16T06:50:29+09:00 Subject: [ruby-talk:15216] Re: Discussion on new Ruby features This is a multi-part message in MIME format. ------=_NextPart_000_000C_01C0DD99.D60EC1C0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable matz wrote: > You don't have to. But it might change Ruby's nature. It might no > longer be Ruby, but Ruby-masked Sather, maybe. I was afraid of this answer.=20 I think, I was unclear: I don't wish any of the features of Sather. I am completely content with the current features of Ruby. The only thing I would like to achieve is a Ruby which produces faster code than Sather :) (of course, this is impossible, but sometimes=20 an impossible objective is a good objective.) I think, Ruby in its current state has everything that a language needs to programm comfortably. The only thing it could need is: performance. My propositions were aimed at enabling the possibility of the=20 programmer to optimize the performance locally without leaving the=20 language. I think the proposed mechanisms are quite surficial and fit into the language smoothly both phylosophically, both because they don't violate backwards-compatibility and since they are optional. They would not make Ruby to a statically typed language or make a principially compiled language. They would jut open the possibility for compilers to optimize the code locally provided the programmer has given some hints. ------=_NextPart_000_000C_01C0DD99.D60EC1C0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable

matz wrote:

> You don't have to. But it = might change=20 Ruby's nature. It might no
> longer be Ruby, but Ruby-masked = Sather,=20 maybe.

I was afraid of this answer.

I think, I was unclear: I don't wish = any of=20 the features of Sather.

I am completely content with the = current=20 features of Ruby. The

only thing I would like to achieve is a = Ruby=20 which produces faster

code than Sather :) (of course, this is = impossible,=20 but sometimes

an impossible objective is a good=20 objective.)

I think, Ruby in its current state has = everything=20 that a language needs

to programm comfortably. The only = thing it=20 could need is:

performance.

My propositions were aimed at enabling = the=20 possibility of the

programmer to optimize the performance locally without leaving the

language. I think the proposed = mechanisms are=20 quite surficial

and fit into the language smoothly both phylosophically,

both because they don't violate = backwards-compatibility and

since they are optional.

They would not make Ruby to a = statically typed=20 language or

make a principially compiled language. = They would=20 jut open the

possibility for compilers to optimize = the=20 code locally provided

the programmer has given some = hints.

------=_NextPart_000_000C_01C0DD99.D60EC1C0--