On Tue, 03 Nov 2009 22:12:10 -0600, r...@rpw3.org (Rob Warnock) said:
> ... > ;; Open new lexical frame for the arguments & initialize them > (setf *env* (augment-environment > *env* > (make-frame (list #:g1581 #:g1582 #:g1583))))
I, too, have been thinking about this [*], essentially along the outline in one of your earlier posts (but, sadly, with too little time to produce an implementation fast enough), and a conclusion I (think I) reached was that one reason why a special variable should not in fact be used, as *ENV* above, is that otherwise bindings might "leak" outside of the given lexical scope within the same dynamic extent. But that is just hand-waving if this here is not Rhodes (the place to jump)...
[*] implementing local lexical variables without LET, without code walking, without replacements for EVAL and APPLY, and "embedded" in Common Lisp, rather than on top of it, though I do not have a strict, precise formulation of the rules of the game
---Vassil.
-- "Even when the muse is posting on Usenet, Alexander Sergeevich?"
Rob Warnock wrote: > Kenneth Tilton <kentil...@gmail.com> wrote: > +--------------- > | Rob Warnock wrote: > | > The critical bit is that the outermost MY-LET must be able to recognize > | > *all* new lexical levels [scopes] that are opened in its body, and insert > | > such redefinitions to all lexical variables defined in outer scopes [but > | > within the outermost MY-LET] to that they reference the correct level from > | > within the inner scope. Since there are many possible ways to open a new > | > lexical scope in CL, MY-LET must therefore be an essentially full code > | > walker [in order to find all such places]. Q.E.D. > | > > | > SKETCH OF PROOF OF CLAIM#2: I already showed that there is nothing > | > "magic" about lexical variables if you reify the lexical environment > | > as a runtime object [e.g., a list of vectors or a vector of vectors]. > | > All you need to do is "compile away" all references to lexical variables > | > into references into the runtime lexical environment object, a snapshot > | > of which gets stored in closures along with the closures' code body. > | > [And, yes, as noted before, you *do* have to redefine all the LAMBDAs > | > in the MY-LET body as well, but I now believe you *don't* necessarily > | > have to redefine EVAL & APPLY per se.] > | > | OK, but now the language CL is not doing this, one is using CL to > | implement a new language. ie, would all the code written be standard CL > | supported according to the CLHS hence portable to any valid CL? > +---------------
> As Ron points out, this has gotten so bizarre that it really should > be considered just an intellectual exercise at this point, but...
...that is Ron's problem, forever sapping any meaning out of a discussion leaving words as mere hollow shells to be shuffled about. Let's not encourage him in his muddy thought.
, yes,
> I'm talking about an approach for which the transformed/rewritten code > *would* still be valid CL and would thus be portable to any valid CL.
Maybe I set the bar in the wrong place. Yes, I can write a COBOL interpreter in portable CL, no that does not mean CL /as a language/ has has MOVE CORRESPONDING.
On Oct 18, 7:58 am, Tim Bradshaw <t...@cley.com> wrote:
> On 2009-10-18 09:12:17 +0100, Ron Garret <rNOSPA...@flownet.com> said:
> > Tomas: while I appreciate the support, could I ask you (and anyone else > > tempted to weigh on on this on my behalf) to please not feed the troll?
> Ah you see, now we come to another ambiguity. In my experience there > are two approaches to trolls. One (the one you suggest here), is to > ignore them until they starve off and die. The other, which I espouse, > is to feed them until they burst spectacularly, spattering their > innards all over bystanders.
> I agree your approach is generally kinder, but it has far lower > entertainment value.
w_a_x_man wrote: > On Nov 3, 10:50 am, Kenneth Tilton <kentil...@gmail.com> wrote: >> Nicolas Neuss wrote: >>> r...@rpw3.org (Rob Warnock) writes: >>>> I thought I was done with this thread... ;-} ;-} >>> And I thought I would never participate:-) >> I take no small pride in having drugged you into this.
> That ought to be "dragged". Someone with a good sixth-grade > education wouldn't make that mistake.
> Similarly, only a monkey would have named a movie > "Honey, I Shrunk the Kids". Anyone else knows the > word should have been "Shrank".
Duane Rettig wrote: > On Nov 6, 3:24 am, w_a_x_man <w_a_x_...@yahoo.com> wrote: >> On Nov 3, 10:50 am, Kenneth Tilton <kentil...@gmail.com> wrote:
>>> Nicolas Neuss wrote: >>>> r...@rpw3.org (Rob Warnock) writes: >>>>> I thought I was done with this thread... ;-} ;-} >>>> And I thought I would never participate:-) >>> I take no small pride in having drugged you into this. >> That ought to be "dragged". Someone with a good sixth-grade >> education wouldn't make that mistake.
> Mistake? I thought "drugged" was entirely appropriate for a great > joke. It is certainly grammatical.
Glad someone can still appreciate a good malajokism.
Vassil Nikolov wrote: > On Fri, 06 Nov 2009 12:06:13 -0500, Kenneth Tilton <kentil...@gmail.com> said: >> Glad someone can still appreciate a good malajokism.
There ya go. Me, I was thinking they spent $50k on focus groups to decide the title of the movie and "Honey, I Shrank the Kids" came in last. It's a fine line between stupid and clever.*
kt
* Spinal Tap (something less than $50k spent on title)
On the other hand, while you can write (setf (car x) v), (setf car) may be not defined: SETF can have specific knowledge of some places without a need for an external setter.
* Kenneth Tilton <4af59568$0$22536$607ed...@cv.net> : Wrote on Sat, 07 Nov 2009 10:42:32 -0500:
| Pascal J. Bourguignon wrote: |> Kenneth Tilton <kentil...@gmail.com> writes: |> |>> There's no way to funcall a setter! FUNCTION does not take variables! | | Ah, silly me, I looked at fdefinition but didn't Get It.
> ;; Example Use > (defstruct foo a) > (defvar $a (make-foo)) > (funcall (find-or-make-setter 'foo-a) 10 $a) ; => #S(FOO :A 10)
> -- > Madhu
Note that this fails for accessors that take more than one argument:
? (find-or-make-setter 'nth) ;Compiler warnings : ; In an anonymous lambda form: In the call to CCL::%SETNTH with arguments (#:G270 #:G269), ; 2 arguments were provided, but at least 3 are required ; by the current global definition of CCL::%SETNTH #<Anonymous Function #x300041F5D2DF> ? (funcall * '(1 2 3) 1 2)
> Error: Too many arguments in call to #<Anonymous Function #x300041F5D2DF>: > 3 arguments provided, at most 2 accepted.
AFAIK, there is no way to fix this in portable CL because there is no potable way to query a function for its argument list, and so no portable way to construct a complete accessor form from a naked accessor function.
You should beware of using the merriam-webster dictionary as a guide to correctness. It simply catalogues current American usage even when that usage is commonly incorrect. Notice that "shrunk" as past tense is given a a variant (here, a common, but incorrect usage). Consider that "shrink" is isomorphic to "drink" and that one might well hear "I drunk me some moonshine," but not "the Queen Mother drunk some tea," unless from the mouth of Eliza Doolittle before her instruction by Prof. Henry Higgins.
On Sun, 08 Nov 2009 10:47:03 -0500, Raffael Cavallaro wrote: > You should beware of using the merriam-webster dictionary as a guide to > correctness. It simply catalogues current American usage even when that > usage is commonly incorrect. Notice that "shrunk" as past tense is
There is no "correct" or "incorrect" usage for (natural) languages. You can distinguish dialects, traditional and modern variants, etc, but none of them are more "correct" than the other. Language is not math.
The best that dictionaries can do is to present common/current usage and variants, so those who aspire to follow them for various reasons (eg because certain people still like to believe that speaking certain variants signals important things about the speaker) can use them as a guide.
> On Sun, 08 Nov 2009 10:47:03 -0500, Raffael Cavallaro wrote:
>> You should beware of using the merriam-webster dictionary as a guide to >> correctness. It simply catalogues current American usage even when that >> usage is commonly incorrect. Notice that "shrunk" as past tense is
> There is no "correct" or "incorrect" usage for (natural) languages. > You can distinguish dialects, traditional and modern variants, etc, > but none of them are more "correct" than the other. Language is not > math.
> The best that dictionaries can do is to present common/current usage > and variants, so those who aspire to follow them for various reasons > (eg because certain people still like to believe that speaking certain > variants signals important things about the speaker) can use them as a > guide.
I beg to differ slightly.
For one thing, I believe most languages are at least to some degree artificial.
I mean, I just cannot believe that complex schemes such as latin declinations evolved naturaly.
Besides, Edo Nyland's "Linguistic Archaeology" seems to give good arguments in this direction.
But even if it wasn't the case historically, to get an efficient communication with a language, we need to use a common reference, a common dictionary and a common grammar.
Of course, I come from a country where there has been for a long time (ie longer than the USA existed) an entity dedicated to establish the dictionary and grammar of the correct French language.
I think this is a good think, because that means that there is less difference between Molière's language and current French than between Shakespear's language and current English, not to say current American.
As a contrarian example, communist chineses imposed a "simplification" of the ideograms to prevent people to read the older chinese texts.
Of course, no dictionary and no law can impose what people think and speak, but you can agree that it is a good thing to aim toward a stable and common vocabulary and grammar, otherwise you wouldn't be programming in COMMON Lisp, but you would choose one of the precursor (MacLisp, ZetaLisp, LeLisp?) , or deviate toward one of the "dialects" (newlisp?).
Pascal J. Bourguignon wrote: > Tamas K Papp <tkp...@gmail.com> writes:
>> On Sun, 08 Nov 2009 10:47:03 -0500, Raffael Cavallaro wrote:
>>> You should beware of using the merriam-webster dictionary as a guide to >>> correctness. It simply catalogues current American usage even when that >>> usage is commonly incorrect. Notice that "shrunk" as past tense is >> There is no "correct" or "incorrect" usage for (natural) languages. >> You can distinguish dialects, traditional and modern variants, etc, >> but none of them are more "correct" than the other. Language is not >> math.
>> The best that dictionaries can do is to present common/current usage >> and variants, so those who aspire to follow them for various reasons >> (eg because certain people still like to believe that speaking certain >> variants signals important things about the speaker) can use them as a >> guide.
> I beg to differ slightly.
> For one thing, I believe most languages are at least to some degree > artificial.
> I mean, I just cannot believe that complex schemes such as latin > declinations evolved naturaly.
And creationists cannot believe the immune system evolved naturally. You are in dubious company.
> Besides, Edo Nyland's "Linguistic Archaeology" seems to give good > arguments in this direction.
> But even if it wasn't the case historically, to get an efficient > communication with a language, we need to use a common reference, a > common dictionary and a common grammar.
Chomsky is such an idiot. He always argued that there must be a language organ otherwise how could infants learn language so perfectly when everyone around them speaks it imperfectly. Hello?
> Of course, I come from a country where there has been for a long time > (ie longer than the USA existed) an entity dedicated to establish the > dictionary and grammar of the correct French language.
C'est vrai, mais mon professor toujours dit "La peuple determine la langue." Is that a famous French saying?
> I think this is a good think, because that means that there is less > difference between Molière's language and current French than between > Shakespear's language and current English, not to say current American.
Seems more like xenophobia than accomplishment to stop a language from changing (when the change would be lexical capture of English words).
> As a contrarian example, communist chineses imposed a "simplification" > of the ideograms to prevent people to read the older chinese texts.
> Of course, no dictionary and no law can impose what people think and > speak, but you can agree that it is a good thing to aim toward a > stable and common vocabulary and grammar, otherwise you wouldn't be > programming in COMMON Lisp, but you would choose one of the precursor > (MacLisp, ZetaLisp, LeLisp?) , or deviate toward one of the "dialects" > (newlisp?).
Proving once again the pointlessnes of argument from analogy.
On Sun, 8 Nov 2009 10:47:03 -0500, Raffael Cavallaro <raffaelcavall...@pas.espam.s.il.vous.plait.mac.com> said:
> On 2009-11-07 01:16:43 -0500, Vassil Nikolov <vniko...@pobox.com> said: >> (By the way: <http://www.merriam-webster.com/dictionary/shrink> >> (inflected forms).) > You should beware of using the [Merriam-Webster] dictionary as a guide > to correctness.
Thank you so much for your guidance, even though I was actually objecting to the statement that "only a monkey" would have used "shrunk" as the past tense form (instead of "shrank"), which is somewhat different.
Which is the body that issues pronouncements when a hitherto incorrect form becomes correct in the English language, or a hitherto correct form ceases to be so?
There are more words in the mouths of people than are dreamt of in your arbiter of correctness, if I may paraphrase...
---Vassil.
-- "Even when the muse is posting on Usenet, Alexander Sergeevich?"
On 2009-11-08 15:23:50 -0500, Vassil Nikolov <vniko...@pobox.com> said:
> Thank you so much for your guidance, even though I was actually > objecting to the statement that "only a monkey" would have used > "shrunk" as the past tense form (instead of "shrank"), which is > somewhat different.
If you read the examples I gave for "drink" I think you'll find this is not too far off the mark, to the extent that a moonshine-drinking redneck is analogous to a monkey.
> Which is the body that issues pronouncements when a hitherto > incorrect form becomes correct in the English language, or a > hitherto correct form ceases to be so?
Formally? None. Informally? The speech of those who have actually been educated in grammar. People educated in grammar know that there is a difference between the past and the past participle. Those uneducated in grammar commonly make the mistake of using a past participle for the past, so much so that this has become the caricature of uneducated speech:
"I seen him do it with my own eyes" "It shrunk away to nothing" "I drunk so much my head is spinning"
are all examples of this common mistake (yes, Tamas, it is possible for native speakers to make mistakes in grammar)
Where:
"I saw him do it with my own eyes!" or "I have seen him do it with my own eyes!" and, "It shrank away to nothing!" "It has shrunk away to nothing!" and, "I drank so much my head is spinning," "I have drunk so much my head is spinning," are correct.
> There are more words in the mouths of people than are dreamt > of in your arbiter of correctness, if I may paraphrase...
On Sun, 8 Nov 2009 18:59:12 -0500, Raffael Cavallaro <raffaelcavall...@pas.espam.s.il.vous.plait.mac.com> said:
> On 2009-11-08 15:23:50 -0500, Vassil Nikolov <vniko...@pobox.com> said: > ... >> Which is the body that issues pronouncements when a hitherto >> incorrect form becomes correct in the English language, or a >> hitherto correct form ceases to be so? > Formally? None. Informally? The speech of those who have actually been > educated in grammar.
As long as this accounts for language evolution adequately. Maybe it does, but if so, that is far from obvious.
---Vassil.
-- "Even when the muse is posting on Usenet, Alexander Sergeevich?"
|
