Now nothing can stop Arc!
http://smuglispweeny.blogspot.com/2008/02/arccells-its-alive.html
MWUAHAHAHAHHAHAHAHAHHAHA!!!!
kenny
--
http://smuglispweeny.blogspot.com/
http://www.theoryyalgebra.com/
"In the morning, hear the Way;
in the evening, die content!"
-- Confucius
Ken Tilton wrote:
> Now nothing can stop Arc!
>
> http://smuglispweeny.blogspot.com/2008/02/arccells-its-alive.html
>
> MWUAHAHAHAHHAHAHAHAHHAHA!!!!
Congratulations! You've ported Cells from Lisp to ... lisp.
Damien Kick wrote:
> Ken Tilton wrote:
>> Now nothing can stop Arc!
>>
>> http://smuglispweeny.blogspot.com/2008/02/arccells-its-alive.html
>>
>> MWUAHAHAHAHHAHAHAHAHHAHA!!!!
>
> Congratulations! You've ported Cells from Lisp to ... lisp.
Or should I say, "Cngrts! Uve prtd Cls from Lsp to ... lsp."
On 2008-02-18 23:42:08 -0500, Damien Kick <·····@earthlink.net> said:
> Or should I say, "Cngrts! Uve prtd Cls from Lsp to ... lsp."
f u cn rd ths u cn gt a gd jb prgrmng n arc
Raffael Cavallaro wrote:
> On 2008-02-18 23:42:08 -0500, Damien Kick <·····@earthlink.net> said:
>
>> Or should I say, "Cngrts! Uve prtd Cls from Lsp to ... lsp."
>
>
> f u cn rd ths u cn gt a gd jb prgrmng n arc
>
(((((I hope you (and Damien) appreciate the irony of Lispers making
(stupid (and obvious (and completely unfunny (jokes) (out of ignorance)
about another programming language.))))))))))
<g>
kenny
--
http://smuglispweeny.blogspot.com/
http://www.theoryyalgebra.com/
"In the morning, hear the Way;
in the evening, die content!"
-- Confucius
On Feb 19, 3:20 am, Ken Tilton <···········@optonline.net> wrote:
> Raffael Cavallaro wrote:
> > On 2008-02-18 23:42:08 -0500, Damien Kick <·····@earthlink.net> said:
>
> >> Or should I say, "Cngrts! Uve prtd Cls from Lsp to ... lsp."
>
> > f u cn rd ths u cn gt a gd jb prgrmng n arc
>
> (((((I hope you (and Damien) appreciate the irony of Lispers making
> (stupid (and obvious (and completely unfunny (jokes) (out of ignorance)
> about another programming language.))))))))))
>
> <g>
>
> kenny
>
> --http://smuglispweeny.blogspot.com/http://www.theoryyalgebra.com/
>
> "In the morning, hear the Way;
> in the evening, die content!"
> -- Confucius
Aaah, the twilight zone. I donno what is real, and what not. Heelp!
On Feb 19, 2:20 am, Ken Tilton <···········@optonline.net> wrote:
> Raffael Cavallaro wrote:
> > On 2008-02-18 23:42:08 -0500, Damien Kick <·····@earthlink.net> said:
>
> >> Or should I say, "Cngrts! Uve prtd Cls from Lsp to ... lsp."
>
> > f u cn rd ths u cn gt a gd jb prgrmng n arc
>
> (((((I hope you (and Damien) appreciate the irony of Lispers making
> (stupid (and obvious (and completely unfunny (jokes) (out of ignorance)
> about another programming language.))))))))))
>
> <g>
Feh. Calling Arc a new language is like calling pig latin a new
language just because everything has a different spelling. It's not
ignorance, I'm just judging a book by its cover. It actually remember
walking past a Walden Books when _Dianetics_ was first released. The
display was hard to miss. Piles of books in the display case with
this eye catching image of a volcano spewing brighlty colored lava
into the air. So I picked up a copy and took a closer look at it.
"There is a single source of all your problems, stress, unhappiness,
depression and self-doubt. It's called the reactive mind. Dianetics
gets rid of it." That claim set off my bullshit detector and I put
the book down. Perhaps I was rejecting it out of ignorance and fear.
Perhaps if I was more open minded, I would've bothered to read the
whole book and then perhaps scheduled myself to get audited.
Apparently, Arc is going to breath new life into stiff limbs of Lisp,
frost bitten from the ravages of the AI winter. I'll let you sign up
for the audit, though, and wait to see if you wind up jumping up and
down on a sofa, proclaiming your love for Cells. <pause> Oh, wait.
You've been doing that for quite some time already <wink>.
But I'm just some poor shlub stuck slinging C++ for cash. Is it any
wonder I'm a bitter old man trapped in a middle aged paunch with a
vindictive sense of humor?
Funny should should imply judging Lisp by all the parentheses on its
cover, though. The first time I saw Lisp, I loved the way it looked.
I think Paul Graham has written something about being able to spot
good lisp code from the shape of it, much like phrenologist feels the
bumps of a head. The first time I saw Lisp code, it vaguely reminded
me of an e.e. cummings poem <http://tinyurl.com/2laac4>. So I
actually started out liking Lisp for very superficial reasons.
On 2008-02-19 03:20:56 -0500, Ken Tilton <···········@optonline.net> said:
> (((((I hope you (and Damien) appreciate the irony of Lispers making
> (stupid (and obvious (and completely unfunny (jokes) (out of ignorance)
^^^^^^^^^^^^^^^^^^^^^^
> about another programming language.))))))))))
This bit assumes I haven't looked at it and used it - I have. It's
principal value appears to be exactly what Ron said it is - the person
behind it has enough popularity to rally a community around it.
The language itself seems in all important respects (unicode, macro
hygiene, modules/libraries) to be more poorly thought out than the
scheme implementation on which it is built, and in some cases (e.g.,
macros), gratuitously broken.
Raffael Cavallaro wrote:
> On 2008-02-19 03:20:56 -0500, Ken Tilton <···········@optonline.net> said:
>
>> (((((I hope you (and Damien) appreciate the irony of Lispers making
>> (stupid (and obvious (and completely unfunny (jokes) (out of ignorance)
>
> ^^^^^^^^^^^^^^^^^^^^^^
>
>> about another programming language.))))))))))
>
>
> This bit assumes I haven't looked at it and used it - I have. It's
> principal value appears to be exactly what Ron said it is - the person
> behind it has enough popularity to rally a community around it.
>
> The language itself seems in all important respects (unicode, macro
> hygiene, modules/libraries) to be more poorly thought out than the
> scheme implementation on which it is built, and in some cases (e.g.,
> macros), gratuitously broken.
>
You don't like CL-style defmacro?!
kenny
--
http://smuglispweeny.blogspot.com/
http://www.theoryyalgebra.com/
"In the morning, hear the Way;
in the evening, die content!"
-- Confucius
On 2008-02-20 19:30:03 -0500, Ken Tilton <···········@optonline.net> said:
> You don't like CL-style defmacro?!
The point is that with *currently available* scheme macro systems you
can have defmacro-style macros (meaning backquote macros) that are
hygienic by default, but allow capture where desired. IOW, the best of
both worlds - write macros in exactly the same style but get hygiene
for free, and the ability to break it in a controlled way as desired.
This is available for mzscheme, the language on which pg is building,
and versions have been available for mzscheme for quite some time now.
To build a hundred year language with capture by default just to get
backquote style macros seems very short-sighted.
Raffael Cavallaro wrote:
> On 2008-02-20 19:30:03 -0500, Ken Tilton <···········@optonline.net> said:
>
>> You don't like CL-style defmacro?!
>
>
> The point is that with *currently available* scheme macro systems you
> can have defmacro-style macros (meaning backquote macros) that are
> hygienic by default, but allow capture where desired.
I have to admit I am not smart like you guys, I could not follow the
thread at all. Was the capture as easy as Just Making the reference to a
variable, or did this require one of those trivial twenty-line
expansions beginning with five left parens?
> IOW, the best of
> both worlds - write macros in exactly the same style but get hygiene for
> free, and the ability to break it in a controlled way as desired.
>
> This is available for mzscheme, the language on which pg is building,
> and versions have been available for mzscheme for quite some time now.
> To build a hundred year language with capture by default just to get
> backquote style macros seems very short-sighted.
>
Just because you and other Schematics want capture to be an issue does
not make it an issue.
As for your assertion that pg did Arc just to get CL's defmacro, that
mess on your screen is a backtrace.
peace. out. k
--
http://smuglispweeny.blogspot.com/
http://www.theoryyalgebra.com/
"In the morning, hear the Way;
in the evening, die content!"
-- Confucius
On 2008-02-21 00:17:56 -0500, Ken Tilton <···········@optonline.net> said:
> I have to admit I am not smart like you guys, I could not follow the
> thread at all. Was the capture as easy as Just Making the reference to
> a variable, or did this require one of those trivial twenty-line
> expansions beginning with five left parens?
It was relatively simple, and moreover, we're talking about a language
with macros here, so you can give it any syntax you like, and certainly
a language designer can give his 'mac' form any syntax he likes.
What matters is the *semantics* and the semantics of cl-style defmacro
is "capture by default, referential transparency uh, not so much." The
semantics of scheme macros is "hygiene and referential transparency by
default, capture when you want it - oh, and btw, you can use backquote
style if you like too."
IOW, the semantics of scheme macros has, as Pascal Costanza has been
pointing out in that other thread, since syntax-case, been strictly
more expressive than cl defmacro. The only win for cl defmacro has been
the syntax of the macro defining form itself - i.e., the ability to use
backquote style macro definitions. But that is no longer the case,
because you can now get scheme semantics *and* backquote style macros
in schmeme.
So, espcially as pg is doing a lisp-1 without a module system there's
no good reason to have non-referentially transparent macros, especially
when the tools to do cl-style backquote macros with scheme semantics
have already been available for mzscheme for a while now (a couple of
years? Andre could give a definitive timeline)
>
>> IOW, the best of both worlds - write macros in exactly the same style
>> but get hygiene for free, and the ability to break it in a controlled
>> way as desired.
>>
>> This is available for mzscheme, the language on which pg is building,
>> and versions have been available for mzscheme for quite some time now.
>> To build a hundred year language with capture by default just to get
>> backquote style macros seems very short-sighted.
>>
>
> Just because you and other Schematics want capture to be an issue does
> not make it an issue.
The CL argument has always been "that's not an issue for us because we
are a lisp-2 and we have packages." Graham can't make that argument.
>
> As for your assertion that pg did Arc just to get CL's defmacro, that
> mess on your screen is a backtrace.
No, re-read what I wrote - he appears to have given 'mac' cl-defmacro
*semantics* just to get backquote style macros, and it turns out that
this is a completely unnecessary trade-off.
> It was relatively simple, and moreover, we're talking about a language
> with macros here, so you can give it any syntax you like, and certainly
> a language designer can give his 'mac' form any syntax he likes.
[snip]
> So, espcially as pg is doing a lisp-1 without a module system there's
> no good reason to have non-referentially transparent macros, especially
> when the tools to do cl-style backquote macros with scheme semantics
> have already been available for mzscheme for a while now (a couple of
> years? Andre could give a definitive timeline)
pg isn't doing a lisp-1 without a modules system, he is doing a lisp-1
whose module system has not yet been implemented. Do not confuse the
two. It is relatively simple, moreover, to keep yourself from
clobbering a function name that is used in a macro. Surely someone who
could understand that mess has the capacity to think before they go
overwrite function names.
> >> IOW, the best of both worlds - write macros in exactly the same style
> >> but get hygiene for free, and the ability to break it in a controlled
> >> way as desired.
>
> >> This is available for mzscheme, the language on which pg is building,
> >> and versions have been available for mzscheme for quite some time now.
> >> To build a hundred year language with capture by default just to get
> >> backquote style macros seems very short-sighted.
>
> > Just because you and other Schematics want capture to be an issue does
> > not make it an issue.
>
> The CL argument has always been "that's not an issue for us because we
> are a lisp-2 and we have packages." Graham can't make that argument.
I think Arc proves it is a non-issue even for lisp-1. Have you tried
it yet? I have never had a problem with this, though everything I've
done so far have been little more than toys so it's not really a good
argument. I don't think you should complain about this yet until
you've actually run into a situation where it was a problem.
As was before stated: just because you are using this as an excuse to
dislike arc, it doesn't mean it is actually an issue.
> > As for your assertion that pg did Arc just to get CL's defmacro, that
> > mess on your screen is a backtrace.
>
> No, re-read what I wrote - he appears to have given 'mac' cl-defmacro
> *semantics* just to get backquote style macros, and it turns out that
> this is a completely unnecessary trade-off.
Could have also used mac like cl-defmacro because he has gone on
record as liking it better. He even wrote a book about it, if you'll
remember?
Consider this: Is there any code you have ever written in common lisp
which couldn't have been trivially implemented in a lisp-1 (I am not
referring to scheme, but rather a theoretical language, whose only
difference from common-lisp is that it is a lisp-1)? I don't believe
so. Even if you did have name conflicts, all you'd have to do is
change the name. That's trivial. But maybe you can come up with
something I can't.
(remember, no using a package system as an excuse to hate arc either,
it will have one, it just hasn't been implemented yet.)
···········@gmail.com wrote:
>> It was relatively simple, and moreover, we're talking about a language
>> with macros here, so you can give it any syntax you like, and certainly
>> a language designer can give his 'mac' form any syntax he likes.
>
> [snip]
>
>> So, espcially as pg is doing a lisp-1 without a module system there's
>> no good reason to have non-referentially transparent macros, especially
>> when the tools to do cl-style backquote macros with scheme semantics
>> have already been available for mzscheme for a while now (a couple of
>> years? Andre could give a definitive timeline)
>
> pg isn't doing a lisp-1 without a modules system, he is doing a lisp-1
> whose module system has not yet been implemented. Do not confuse the
> two. It is relatively simple, moreover, to keep yourself from
> clobbering a function name that is used in a macro. Surely someone who
> could understand that mess has the capacity to think before they go
> overwrite function names.
According to
http://library.readscheme.org/servlets/cite.ss?pattern=Rees-AIM-Baw-88
accidental variable capture seems to happen quite often in a Lisp-1.
> I think Arc proves it is a non-issue even for lisp-1. Have you tried
> it yet? I have never had a problem with this, though everything I've
> done so far have been little more than toys so it's not really a good
> argument. I don't think you should complain about this yet until
> you've actually run into a situation where it was a problem.
I think it's too early to tell what Arc does and doesn't prove.
Pascal
--
1st European Lisp Symposium (ELS'08)
http://prog.vub.ac.be/~pcostanza/els08/
My website: http://p-cos.net
Common Lisp Document Repository: http://cdr.eurolisp.org
Closer to MOP & ContextL: http://common-lisp.net/project/closer/
Damien Kick wrote:
> Damien Kick wrote:
>
>> Ken Tilton wrote:
>>
>>> Now nothing can stop Arc!
>>>
>>> http://smuglispweeny.blogspot.com/2008/02/arccells-its-alive.html
>>>
>>> MWUAHAHAHAHHAHAHAHAHHAHA!!!!
>>
>>
>> Congratulations! You've ported Cells from Lisp to ... lisp.
Ah, if you were any smarter you would realize it was a port between
communities, one dead, one newborn.
> Or should I say, "Cngrts! Uve prtd Cls from Lsp to ... lsp."
Dying is easy, comedy is hard. And you wonder why you are in my killfile!
peace, kenny
--
http://smuglispweeny.blogspot.com/
http://www.theoryyalgebra.com/
"In the morning, hear the Way;
in the evening, die content!"
-- Confucius
On Feb 19, 9:23 am, Ken Tilton <···········@optonline.net> wrote:
> Damien Kick wrote:
> >> Congratulations! You've ported Cells from Lisp to ... lisp.
>
> Ah, if you were any smarter you would realize it was a port between
> communities, one dead, one newborn.
My Kennyquites dictionary just got a new addition, thanks to both of
you.
Slobodan
On Feb 19, 2:23 am, Ken Tilton <···········@optonline.net> wrote:
> Damien Kick wrote:
> > Damien Kick wrote:
>
> >> Ken Tilton wrote:
>
> >>> Now nothing can stop Arc!
>
> >>>http://smuglispweeny.blogspot.com/2008/02/arccells-its-alive.html
>
> >>> MWUAHAHAHAHHAHAHAHAHHAHA!!!!
>
> >> Congratulations! You've ported Cells from Lisp to ... lisp.
>
> Ah, if you were any smarter you would realize it was a port between
> communities, one dead, one newborn.
Well, if you're dropping by again, do pop in, huh? And thanks a lot
for the
gold, and frankincense, but don't worry too much about the myrrh next
time,
all right? Thank you. Goodbye! ... Well, weren't they nice? Out of
their
bloody minds, still...
> > Or should I say, "Cngrts! Uve prtd Cls from Lsp to ... lsp."
>
> Dying is easy, comedy is hard. And you wonder why you are in my killfile!
Damiecus: Boy, when you die at the palace, you really *die* at the
palace.
Chris Barts wrote:
> Damien Kick <·····@earthlink.net> writes:
>
>
>>Congratulations! You've ported Cells from Lisp to ... lisp.
>
>
> Yep, and Java and C++ are both Algol.
Funny you should mention that, I have done ports to Python, Java, C++,
and two databases as well, AllegroStore and AllegroGraph. I love
watching databases update themselves.
I think the coolest thing about this is that a child of two could look
at that blog entry and "get it", at least on the implementation angle.
But it is not quite Cells3, in between actually, Cells2.5.
kenny
--
http://smuglispweeny.blogspot.com/
http://www.theoryyalgebra.com/
"In the morning, hear the Way;
in the evening, die content!"
-- Confucius
Ken Tilton <···········@optonline.net> writes:
> Chris Barts wrote:
>> Damien Kick <·····@earthlink.net> writes:
>>
>>
>>>Congratulations! You've ported Cells from Lisp to ... lisp.
>>
>>
>> Yep, and Java and C++ are both Algol.
>
> Funny you should mention that, I have done ports to Python, Java, C++,
> and two databases as well, AllegroStore and AllegroGraph. I love
> watching databases update themselves.
"IT'S ALIVE! IT'S ALIVE! NOW I KNOW HOW IT FEELS TO BE GOD!"
>
> I think the coolest thing about this is that a child of two could look
> at that blog entry and "get it", at least on the implementation
> angle.
At least his patience wouldn't be tried by long names.
Ken Tilton wrote:
> Now nothing can stop Arc!
>
> http://smuglispweeny.blogspot.com/2008/02/arccells-its-alive.html
>
> MWUAHAHAHAHHAHAHAHAHHAHA!!!!
I forgot to taunt: Cells/Arc is better than Cells/CL. One can define
observers to go along with one rule, to go along with one slot of any
instance of a rule, or to apply globally to all slots of a given name.
Eat your hearts out!
:)
kenny
--
http://smuglispweeny.blogspot.com/
http://www.theoryyalgebra.com/
"In the morning, hear the Way;
in the evening, die content!"
-- Confucius
Ken Tilton <·········@gmail.com> wrote:
+---------------
| I forgot to taunt: Cells/Arc is better than Cells/CL. One can define
| observers to go along with one rule, to go along with one slot of any
| instance of a rule, or to apply globally to all slots of a given name.
+---------------
But how much of that comes from having done Cells/CL *first*?
And if you re-did Cells/CL again from scratch, knowing what
you now know about original Cells/CL and Cells/Arc, wouldn't
Cells/CL The Re-Write be even better than either?!?
-Rob
-----
Rob Warnock <····@rpw3.org>
627 26th Avenue <URL:http://rpw3.org/>
San Mateo, CA 94403 (650)572-2607
Rob Warnock wrote:
> Ken Tilton <·········@gmail.com> wrote:
> +---------------
> | I forgot to taunt: Cells/Arc is better than Cells/CL. One can define
> | observers to go along with one rule, to go along with one slot of any
> | instance of a rule, or to apply globally to all slots of a given name.
> +---------------
>
> But how much of that comes from having done Cells/CL *first*?
Aw, I am just abusing the yobbos, I was not making a serious point.
In fact, this greater observer flexibility may be a solution in search
of a problem: Cells has always been driven by necessity, and necessity
has never asked for anything other than what is there now, a GF
specified on the instance, the slot-name, the new-value, and the
old-value, and 95% of the time the slot-name decides all.
But Arc makes programming so damn easy I just said, why not?, and tossed
it off in a few minutes.
Hmmm, just thought of one use: debugging, tho I guess the same effect
could be had from within the rule.
>
> And if you re-did Cells/CL again from scratch, knowing what
> you now know about original Cells/CL and Cells/Arc, wouldn't
> Cells/CL The Re-Write be even better than either?!?
No, the way I work is to always move towards the light. I am not smart
like you guys, I cannot write great code off the top of my head, I just
know when I have written bad code and then I change it. So if there was
anyway to make Cells better I would be coding it now.
What /would/ make Cells better is a new application. Cells-ODE is in the
works and I will take Cells/Arc and use it to flesh out a little further
TripleCells -- the nice thing about Cells/Arc being that Arc makes
code so much tighter and transparent that I now have a 325-line
implementation of what was 2700 lines of CL. Well, lessee how much
further getting to 3.0 takes it. Oh, and synapses. Ah, forgot the whole
Family class, that is a /lot/ of code...
kenny
--
http://smuglispweeny.blogspot.com/
http://www.theoryyalgebra.com/
"In the morning, hear the Way;
in the evening, die content!"
-- Confucius
On Feb 19, 10:28 am, Ken Tilton <···········@optonline.net> wrote:
> What /would/ make Cells better is a new application. Cells-ODE is in the
> works and I will take Cells/Arc and use it to flesh out a little further
> TripleCells -- the nice thing about Cells/Arc being that Arc makes
> code so much tighter and transparent that I now have a 325-line
> implementation of what was 2700 lines of CL. Well, lessee how much
> further getting to 3.0 takes it. Oh, and synapses. Ah, forgot the whole
> Family class, that is a /lot/ of code...
Not quite sure what not-quite Cells 3.0 is supposed to mean, but it
looks like at-rest values, laziness, and data integrity are all
missing ... which certainly won't prevent it from being useful for the
size of things people seem to be doing in Arc, but given that, I don't
think the line count is very surprising. Add those, and the ability
to supply a function to determine if the value has changed, and in CL
you're talking approx 500-700 LOC.
Actually, it's interesting, I *don't* see that code looking any
shorter than its CL equivalent. The code is more concise as in fewer
characters per line, but you have about the same number of lines, same
number of forms.
Thomas F. Burdick wrote:
> Actually, it's interesting, I *don't* see that code looking any
> shorter than its CL equivalent. The code is more concise as in fewer
> characters per line, but you have about the same number of lines, same
> number of forms.
It's the parens! Count the parens! They are almost all gone!
How is Arc port of C4 going?
hth, kenny
--
http://smuglispweeny.blogspot.com/
http://www.theoryyalgebra.com/
"In the morning, hear the Way;
in the evening, die content!"
-- Confucius
Ken Tilton wrote:
>
>
> Thomas F. Burdick wrote:
>
>> Actually, it's interesting, I *don't* see that code looking any
>> shorter than its CL equivalent. The code is more concise as in fewer
>> characters per line, but you have about the same number of lines, same
>> number of forms.
>
>
> It's the parens! Count the parens! They are almost all gone!
A lot of those, btw, are saved by being able to say sqrt.x instead of
(sqrt x) when there is only one arg. But I am still a noob so I write
(sqrt x) and then later remember the syntax and edit.
Will a light come on? When I am typing, will my fingers learn that I am
starting a one param form and steer clear of the left parens?
Another trick (one I have not used) is (sqrt:sqrt 16)->2. But I cannot
combine those:
sqrt:sqrt.16
...returns a procedure. Wonder what it does.
As for (sqrt:sqrt 16), would my fingers ever learn when I am piping and
go for the : instead of space-(?
I may not find out, fleshing out TripleCells today.
kt
--
http://smuglispweeny.blogspot.com/
http://www.theoryyalgebra.com/
"In the morning, hear the Way;
in the evening, die content!"
-- Confucius
P� Tue, 19 Feb 2008 17:25:33 +0100, skrev Ken Tilton
<···········@optonline.net>:
>
>
> A lot of those, btw, are saved by being able to say sqrt.x instead of
> (sqrt x) when there is only one arg. But I am still a noob so I write
> (sqrt x) and then later remember the syntax and edit.
>
> Will a light come on? When I am typing, will my fingers learn that I am
> starting a one param form and steer clear of the left parens?
>
> Another trick (one I have not used) is (sqrt:sqrt 16)->2. But I cannot
> combine those:
>
> sqrt:sqrt.16
>
sqrt.3.1415 gives a error.. it is read as (sqrt 3 1415)
(sqrt 3.1415) works..
I fail to see this syntax as a good thing.
Also I can't see if something is declared a function, hash-table, or a
access into a list.
This may be deliberate, but I find it more often confuses me than helps.
(= a (list 1 2 3 4 5 6 7 8 9))
(a 3) ==> 4
(def a (x) (* x x))
; redefinition gives warning
(a 3) ==> 9
(= a (table))
; redefines quietly
(= (a 3) 4)
(a 3) ==> 4
--------------
John Thingstad
On Feb 19, 9:43 am, "John Thingstad" <·······@online.no> wrote:
> I fail to see this syntax as a good thing.
Of course it's better! You're trading a few extra characters for the
chance to not remember the corner cases where the parser doesn't
work! And which do you do more of - typing or figuring out broken
corner cases? Huh?
Of course, it may be another case of "Worse is Worse", but who am I to
say?
faa
Ken Tilton wrote:
>
>
> Rob Warnock wrote:
>> Ken Tilton <·········@gmail.com> wrote:
>> +---------------
>> | I forgot to taunt: Cells/Arc is better than Cells/CL. One can define
>> | observers to go along with one rule, to go along with one slot of
>> any | instance of a rule, or to apply globally to all slots of a given
>> name.
>> +---------------
>>
>> But how much of that comes from having done Cells/CL *first*?
>
> Aw, I am just abusing the yobbos, I was not making a serious point.
>
> In fact, this greater observer flexibility may be a solution in search
> of a problem: Cells has always been driven by necessity, and necessity
> has never asked for anything other than what is there now, a GF
> specified on the instance, the slot-name, the new-value, and the
> old-value, and 95% of the time the slot-name decides all.
>
> But Arc makes programming so damn easy I just said, why not?, and tossed
> it off in a few minutes.
There is a bug in the dynamic scoping construct. Does Arc have something
like unwind-protect?
Pascal
--
1st European Lisp Symposium (ELS'08)
http://prog.vub.ac.be/~pcostanza/els08/
My website: http://p-cos.net
Common Lisp Document Repository: http://cdr.eurolisp.org
Closer to MOP & ContextL: http://common-lisp.net/project/closer/