Re: [9fans] more extensions

Re: [9fans] more extensions

[bwc]

[-- Attachment #1: Type: text/plain, Size: 109 bytes --]

I'll amen Howard here.  I'm just now getting used to function prototypes!

(that was a joke)

 Brantley

[-- Attachment #2: Type: message/rfc822, Size: 2014 bytes --]

From: "Howard Trickey" <howard@research.bell-labs.com>
To: <9fans@cse.psu.edu>
Subject: [9fans] more extensions
Date: Tue, 16 Jul 2002 13:28:07 -0400
Message-ID: <000501c22cee$2608e170$bf356887@bl.belllabs.com>

This is tempting.  Tuples are one of my favorite things
about Limbo.

However, I'm biased, as I was the one that pushed for tuples
in Limbo, I think.

But adding them to C's already delicate grammar might
be harder than it appears.  And  there comes a point
where an extension changes the flavor of a language
a little too much, and I think this one might be over
that line.

- Howard Trickey

> Including tuples (perhaps as an instance of nameless structs) would
> simplify many interfaces, right? that's one thing I learned using
> Limbo.

Re: [9fans] more extensions

[Fariborz (Skip) Tavakkolian]

That reminds me of a million dollar idea I had the other day. As we are all nearing
retirement, it would be interesting to consider retirement homes equipped with
old machines that only understand old dialects of C, FORTRAN, etc.  I'm telling you
the retiring software developer generation is a gold mine. The keyboards would
have to be drool proofed though.

(also a joke)

At 03:15 PM 7/16/2002 -0400, bwc@coraid.com wrote:
>I'll amen Howard here.  I'm just now getting used to function prototypes!
>
>(that was a joke)
>
> Brantley

Re: [9fans] more extensions

[Douglas A. Gwyn]

"rob pike, esq." wrote:
> ... I think C is not the place to do language hacking any more.

I largely agree, although there are some places in C where programmers
really could use some help.  The C language is not very "clean" due
to having accumulated features rather than having features redesigned
(which would of course have caused major problems for important
existing code).  Any really substantial changes would be done better
in a new language.  Maybe language designers could learn from
history and avoid many of the bigger mistakes, mainly overloading
single constructs with several conceptually non-identical meanings.

Re: [9fans] more extensions

[Douglas A. Gwyn]

FJ Ballesteros wrote:
> Including tuples ...

Tuples are truly useful.  There are some syntactic issues to solve
when adding them to C, but assuming they're dealt with I would
welcome this experiment.  It would provide some leverage in arguing
for standardization of the feature for C0x.

Re: [9fans] more extensions

[Jon Snader]

On Tue, Jul 16, 2002 at 02:47:27PM -0400, Dan Cross wrote:
> 
> Not in response to Scott's comment, but in general, I'm not a big fan
> of changing C; C is what C is.  If you want something else, use
> something else, but don't try to morph C into another language.
> 

As a wise man once said (sort of): ``If you want limbo, you know
where to get it.''

Re: [9fans] more extensions

[Dan Cross]

> | Including tuples (perhaps as an instance of nameless structs) would
> | simplify many interfaces, right? that's one thing I learned using
> | Limbo.
> 
> Yeah, but limbo has bounds checked arrays, which are even more useful.

Has anyone ever wondered, ``why not just port limbo?''  Surely the
runtime environment provided by Dis can be emulated using a library.

Not in response to Scott's comment, but in general, I'm not a big fan
of changing C; C is what C is.  If you want something else, use
something else, but don't try to morph C into another language.

	- Dan C.

Re: [9fans] more extensions

[rob pike, esq.]

[-- Attachment #1: Type: text/plain, Size: 311 bytes --]

I love tuples, just love them, but I think C is not the place to
do language hacking any more.  A case could even be made
that it wasn't the right place when Ken hacked it around 1988,
although a number of the changes were valuable and did make
it into the standard (e.g. indexed initializations).

-rob

[-- Attachment #2: Type: message/rfc822, Size: 2088 bytes --]

From: FJ Ballesteros <nemo@gsyc.escet.urjc.es>
To: 9fans@cse.psu.edu
Subject: [9fans] more extensions
Date: Tue, 16 Jul 2002 19:24:58 +0200
Message-ID: <3D3456EA.4B0BC1D4@gsyc.escet.urjc.es>


Including tuples (perhaps as an instance of nameless structs) would
simplify many interfaces, right? that's one thing I learned using
Limbo.

The extension would probably be not too complex, also. 

I don't have time to implement this during this semester, but would
be willing to do so in the future if people from the Labs agree.

Is there agreement on this?

Re: [9fans] more extensions

[Scott Schwartz]

| Including tuples (perhaps as an instance of nameless structs) would
| simplify many interfaces, right? that's one thing I learned using
| Limbo.

Yeah, but limbo has bounds checked arrays, which are even more useful.

[9fans] more extensions

[Howard Trickey]

This is tempting.  Tuples are one of my favorite things
about Limbo.

However, I'm biased, as I was the one that pushed for tuples
in Limbo, I think.

But adding them to C's already delicate grammar might
be harder than it appears.  And  there comes a point
where an extension changes the flavor of a language
a little too much, and I think this one might be over
that line.

- Howard Trickey

> Including tuples (perhaps as an instance of nameless structs) would
> simplify many interfaces, right? that's one thing I learned using
> Limbo.

[9fans] more extensions

[FJ Ballesteros]

Including tuples (perhaps as an instance of nameless structs) would
simplify many interfaces, right? that's one thing I learned using
Limbo.

The extension would probably be not too complex, also. 

I don't have time to implement this during this semester, but would
be willing to do so in the future if people from the Labs agree.

Is there agreement on this?

Re: [9fans] more extensions

[Sam]

My next thought, but I figured I'd used all my contribution
credits today.

Sam

On Tue, 16 Jul
2002, FJ Ballesteros wrote:

>
> Including tuples (perhaps as an instance of nameless structs) would
> simplify many interfaces, right? that's one thing I learned using
> Limbo.
>
> The extension would probably be not too complex, also.
>
> I don't have time to implement this during this semester, but would
> be willing to do so in the future if people from the Labs agree.
>
> Is there agreement on this?
>