 |
|
| Science Forum Index » Languages Forum » All languages are equally fit... |
|
Page 62 of 83 Goto page Previous 1, 2, 3 ... 61, 62, 63 ... 81, 82, 83 Next |
|
| Author |
Message |
| António Marques... |
 Posted: Sat Nov 07, 2009 1:50 pm |
|
|
|
Guest
|
On Nov 7, 5:54 am, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
[quote]António Marques (in sci.lang):
On Nov 7, 1:08 am, LEE Sau Dan <dan... at (no spam) informatik.uni-freiburg.de
wrote:
"Nathan" == Nathan Sanders <nsand... at (no spam) williams.edu> writes:
There is an example where ambiguity is _not_ ungrammatical but by
design. In the database language SQL, you can write
select * from A where x > y
(Assuming that you have a table named A that has two number
columns x and y.) Now this will produce a list of all rows where
the x entry is bigger than the y entry, but the order in which
the rows appear is unpredictable. (The database system will use
table statistics, recent buffered query results and whatever to
decide which order it can produce in the shortest time).
Nathan> "And whatever"? Is it really undefined and random?
Yes. The spec. says clearly that the order of the results rows are in
an unspecified order. i.e. you as a programmer should not make any
assumption on the ordering of the resulting rows. An implementation is
allowed to use any order, and the order does not have to be consistent a
second time you use an identical query.
The order is arbitrary, unpredictable. Whether it is random is up to
the implementation. (Not that "random" is just one possibility to
realize "undefined", and it differs from "arbitrary".)
Is that hard to understand? Is that ambiguous?
Sorry, but where is the ambiguity? Of course the results of a SELECT
query are unordered, unless there is an ORDER BY clause. What's
ambiguous about that? Where does that enter the discussion?
It is ambiguous in so far as the outcome is not predictable.
[/quote]
The outcome is predictable, since the outcome is known to be
unordered. So, 'A B C D' and 'A C D B' are the 'same' outcome. Only
the properties expected (and here order is not one of them) are
relevant.
(I actually should know; I routinely spend time reading database rows
into objects and putting those into hashtables - which, again, usually
have no ordering - and it's all the same whether A or B comes first,
either in the original row set or in the resulting hashtable. Whenever
order matters, the query must be different and the hashtable must
present itself sorted by order of insertion.) |
|
|
| Back to top |
|
|
|
| António Marques... |
| Posted: Sat Nov 07, 2009 2:00 pm |
|
|
|
Guest
|
On Nov 7, 9:08 am, "PaulJK" <paul.kr... at (no spam) paradise.net.nz> wrote:
[quote]There is nothing difficult to understand about "select * from A where x > y"
and there is only one way one can interpret it, simply the statement
instructs SQL to (...)
[/quote]
When I saw SQL mentioned I thought it was because of its dauntingly
hard grammar. The fact is that there you have a language which is so
complex that true ambiguity arises at every corner. But what precisely
happens is that every implementation solves the ambiguities - its own
way, but it does solve them. |
|
|
| Back to top |
|
|
|
| DKleinecke... |
| Posted: Sat Nov 07, 2009 3:37 pm |
|
|
|
Guest
|
On Nov 7, 7:32 am, hru... at (no spam) odds.stat.purdue.edu (Herman Rubin) wrote:
[quote]In article <afe54dcf-aa3a-4076-bbfe-de821b21a... at (no spam) d21g2000yqn.googlegroups..com>,
=?ISO-8859-1?Q?Ant=F3nio_Marques?= <ento... at (no spam) gmail.com> wrote:
On Nov 6, 9:19=A0pm, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
Ant=F3nio Marques (in sci.lang):
On 6 Nov, 18:44, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
Why do you think higher level
programming languages exist? They have been invented because it is eas> >ier
to write complex programs in them than it is if you write machine code
that does the equivalent job. Some programming languages are flexible > >in
the code they accept precisely to allow for pragmatics in the code.
Machine code, yes. I believe that a properly designed
"higher level" assembler code can be almost as concise,
and every bit as easy to write, as the current high
level languages, which are very definitely weak in many
places.
--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Department of Statistics, Purdue University
hru... at (no spam) stat.purdue.edu Phone: (765)494-6054 FAX: (765)494-0558
[/quote]
Assuming that by assembly code you mean code that mirrors a computer's
internal instruction set this is a reasonable opinion. But the
results are computer-dependent. The assembly code must be rewritten
for each different species of computer. I do not know how many
versions would required today but at least a half dozen would be
needed. People who aim to code for all computers - like the GNU
project - simply cannot afford that much investment. A business like
Microsoft can cope because it is a near monopoly and simply ignores
all but a few kinds of computers.
Modern compilers work in three (at least steps). The first step reads
the higher-order language and compiles it into an intermediate
language. The second step optimizes the code in the intermediate
language and the third step realizes that language in the internal
instruction set of the target computer. Assembly language refers to
either the intermediate language or the final instruction set. If you
want to all you have to do is interrupt this process after the second
step and meddle with the code.
There are good reasons why nobody actually does this. The assembly
language versions the compiler presents are too hard to understand.
Perhaps better (equals more comprehensible) assembly languages can be
invented (I have done this and decided that nobody cared). But the
real barrier is comprehension. Almost everything done on computers
these days in done in a fashion that almost defeats human
comprehension even when it not a secret. Have you ever tried to
understand the Linux kernel? Or the Gnome window manager? And these
aren't even application programs.
Perhaps someday we will have few enough different kinds of computers
(one would be optimal) and good enough current programs that we can
afford to use assembly language. I don't expect it in my lifetime. |
|
|
| Back to top |
|
|
|
| DKleinecke... |
| Posted: Sat Nov 07, 2009 3:39 pm |
|
|
|
Guest
|
On Nov 7, 7:29 am, hru... at (no spam) odds.stat.purdue.edu (Herman Rubin) wrote:
[quote]In article <87vdhni2be.... at (no spam) informatik.uni-freiburg.de>,
LEE Sau Dan <dan... at (no spam) informatik.uni-freiburg.de> wrote:
"Nathan" =3D=3D Nathan Sanders <nsand... at (no spam) williams.edu> writes:
.......................
What level of "interpretation" do you want?
Abstraction level is important!
If you ignore abstraction level, then perhaps you shouldn't intepret "0"
and "1" as merely "zero" and "one". Strictly speaking, you should dig
deeping into the definition of them. E.g. Richard Dedekind defines "0"
as the empty set {}, and "1" as the set {0} =3D {{}}.. 2 =3D {0,1} > >=3D
{{},{{}}}, etc.
That is NOT Dedekind's representation. It is von Neumann's.
Dedekind had 2 = {1}, etc.
So, do you ALWAYS want to think about "0" and "1" that way?
Dedekind clearly stated that if it looks like the integers,
and acts like the integers, it is a representation of the
integers. Except for rather esoteric situations, it makes
no difference which representation is used.
One of the problems in teaching the conceptS of the integers
is that there are many, and the two most "basic" ones appear
to have nothing in commone Nevertheless, the operations on
them are the same, and one can use either of these, and others,
and one is still using the integers.
--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Department of Statistics, Purdue University
hru... at (no spam) stat.purdue.edu Phone: (765)494-6054 FAX: (765)494-0558
[/quote]
Aren't you just saying "use the axiomatic approach"? |
|
|
| Back to top |
|
|
|
| DKleinecke... |
| Posted: Sat Nov 07, 2009 3:41 pm |
|
|
|
Guest
|
On Nov 7, 8:24 am, "Peter T. Daniels" <gramma... at (no spam) verizon.net> wrote:
[quote]On Nov 7, 10:14 am, hru... at (no spam) odds.stat.purdue.edu (Herman Rubin) wrote:
In article <8e8a3450-0054-408b-9aac-d6badc252... at (no spam) u20g2000vbq.googlegroups.com>,
Peter T. Daniels <gramma... at (no spam) verizon.net> wrote:
On Nov 6, 12:57=A0pm, hru... at (no spam) odds.stat.purdue.edu (Herman Rubin) wrote:
In article <871a229f-73b6-43e9-9c34-efebf3392... at (no spam) d5g2000yqm.googlegroups.> > >com>,
Peter T. Daniels <gramma... at (no spam) verizon.net> wrote:
On Nov 5, 7:13=3DA0pm, hru... at (no spam) odds.stat.purdue.edu (Herman Rubin) wrote:
In article <1c95f087-8202-468b-852a-ad1019017... at (no spam) g27g2000yqn.googlegro> > >ups=3D
.com>,
Peter T. Daniels <gramma... at (no spam) verizon.net> wrote:
On Nov 5, 8:31 am, LEE Sau Dan <dan... at (no spam) informatik.uni-freiburg.de
wrote:
.....................
What weapon can be made using dynamite? =3DA0Only a bomb.
In cartoons and cheap TV shows, maybe. There are much better ways to
make bombs!
There are now, but in the time of Nobel, dynamite was very
much used for the purpose; the newer explosives came later.
My comment still holds; what other weapon is made using
dynamite? =A0
Maybe it depends on your definition of "bomb." Were there no bombs
before Nobel?
Nitroglycerin is an explosive which is hard to handle
and can go off unexpectedly. Dynamite made it safer.
Eh? Bombs were made with nitroglycerin until late in the 19th century?
Maybe it depends on your definition of "bomb."
[/quote]
When Clint Eastwood used dynamite it was more like a hand grenade. |
|
|
| Back to top |
|
|
|
| DKleinecke... |
| Posted: Sat Nov 07, 2009 3:50 pm |
|
|
|
Guest
|
On Nov 7, 1:41 am, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
[quote]Nathan Sanders (in sci.lang):
In article <877hu2g7f0.... at (no spam) informatik.uni-freiburg.de>,
LEE Sau Dan <dan... at (no spam) informatik.uni-freiburg.de> wrote:
"Nathan" == Nathan Sanders <nsand... at (no spam) williams.edu> writes:
>> Maybe, you could try using the postfix notation instead to see if
>> that helps stopping them from thinking so. When people see
>> something unfamiliar (postfix notation in this case), they may
>> refrain from generalizing their past knowledge into the new
>> field.
Nathan> Why would I introduce a completely different notation from
Nathan> that that is (a) most common and (b) used in the textbook?
Not a completely different one. But you may need to use non-standard
things for:
1) new concepts;
Actually, we do cover concepts not formalized in the textbook
(pragmatics), and I do introduce new formal notation for those
concepts. However, this comes near the end of the semester, after
they're already long-accustomed to the standard notation in the
textbook (infixed binary logical connectives and prefixed quantifiers,
negation, and predicates), so I'm not going to switch things up on
them in the last few weeks. Especially not with postfix notation!
2) convenience (to introduce short-hands, such as Einstein's notation)
The point of the formal definitions is to have precision. Convenience
is nice, but is not the primary goal (or even a goal; it doesn't
really matter at all how "convenient" the logical system is).
I hated the kind of mathematics that depends on the correctness of long
algebraic calculations. If I have to believe a fact that is only founded on
manipulation of formulas, I don't get the insight why it is true, and so
this is error-prone. I preferred arriving at results by insights and using
formulas only as a second step putting those insights on paper, hopefully
invoking the insight in the reader. Normally, prose works better for that
purpose.
[/quote]
Unless you have an insight into the distribution of prime numbers I
don't see how you come to believe any theorem about the distribution
of prime numbers. If you have an insight into the distribution of
prime numbers please share it with us. |
|
|
| Back to top |
|
|
|
| Joachim Pense... |
| Posted: Sat Nov 07, 2009 5:45 pm |
|
|
|
Guest
|
Peter T. Daniels (in sci.lang):
[quote]On Nov 7, 4:35Â am, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
Nathan Sanders (in sci.lang):
In article <87fx8rgg7m.... at (no spam) gmail.com>,
LEE Sau Dan <dan... at (no spam) informatik.uni-freiburg.de> wrote:
"Nathan" == Nathan Sanders <nsand... at (no spam) williams.edu> writes:
No. You seem to consider mathematical notation as a
prescriptivist considers English.
Nathan> Of course. Â Mathematical notation, unlike language, was
Nathan> consciously and intentionally designed.
Designed by whom? Â God?
By humans, obviously.
If you  realized that it  is designed by  human beings, would  you
still suppose that it is flawless?
I don't "suppose" it; I know it.
Are you really unaware that mathematical logic defines the notion of
well-formed formula?
Can humans design flawed notation systems? Â Of course. Â Those are
broken systems that good mathematicians would not use, because such
systems result in exactly the kind of ambiguity and imprecision that
mathematical notation is ordinarily designed to avoid. Â How terrible
it would be to have a mathematical notation with no defined order of
operations or grouping!
Why is ambiguity in language more acceptable?
Because language occurs in a social context that includes a great deal
of information that's not explicit in the utterance. And because
utterances carry presuppositions (and other properties), as shown by
the "only ... but" examples.
[/quote]
So you are not talking about real ambiguity but about pseudo-ambiguity that
occurs when you consider utterances isolated from their context?
I don't see a difference here between formulas, language, programming: If
you allow ambiguity that is neither eliminated by context nor irrelevant in
the context, then your communication (or programming) attempt fails.
(Except in those cases in poetry or jokes where you deliberately exploit
ambiguity).
Joachim |
|
|
| Back to top |
|
|
|
| Nathan Sanders... |
| Posted: Sat Nov 07, 2009 6:08 pm |
|
|
|
Guest
|
In article <hd3q1s$n85$01$1 at (no spam) news.t-online.com>,
Joachim Pense <snob at (no spam) pense-mainz.eu> wrote:
[quote]António Marques (in sci.lang):
On 7 Nov, 06:05, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
António Marques (in sci.lang):
On Nov 6, 9:19Â pm, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
António Marques (in sci.lang):
On 6 Nov, 18:44, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
Why do you think higher level
programming languages exist? They have been invented because it is
easier to write complex programs in them than it is if you write
machine code that does the equivalent job. Some programming
languages are flexible in the code they accept precisely to allow
for pragmatics in the code.
I must be missing something. What exactly (in programming languages)
are you talking about?
Pragmatics. I had given an example before.
Must have missed it. Be so kind as to elaborate, give another example,
etc.
The following lines of Perl all do the same thing but are pragmatically
different:
if ($a != 1) {$b = 17}
unless ($a == 1) {$b = 17}
$b = 17 unless $a == 1
$a == 1 or $b = 17
$a != 1 and $b = 17
I see what you mean. Sorry I hadn't spotted it earlier and made you
repeat it.
However, I don't think these are examples of pragmatics. They are
structurally different ways of getting the exact same result; their
intent, as you mention, is to better model what's on the programmer's
mind - say, make the language more semantic programmer-wise. In what
way do you consider it to be an example of pragmatics?
In my view these are examples of topicalizing a sentence differently. Isn't
topicalization part of pragmatics?
[/quote]
It is, but is it part of programming languages? What observable,
functional role does it play? If the compiler cares, then it's not
pragmatics; if the compiler doesn't care, then why should we?
But even if something like pragmatics plays an identifiable role in
computer programming, is it really an inherent part of the computer
language itself, or just something copied over from the programmer's
own native language? That is, do all typical programmers writing in
the same computer language for the same compilers have identical
pragmatic behavior, or does their pragmatic behavior vary based on
their native language? (Or worse, does it vary from individual to
individual in unpredictable ways? If so, that's not pragmatics.)
[quote]I think "better model what's on the programmer's mind" is more a pragmatic
than a semantic concept.
[/quote]
I remain unconvinced that, in the context of the properties of
computer languages, "what's on the programmer's mind" is a concept
worth making a model of, or even possible to make a useful model of.
Is "what's on the programmer's mind" really consistent and predictable
enough to even be meaningfully modeled?
Nathan
--
Nathan Sanders
Linguistics Program
Williams College
http://wso.williams.edu/~nsanders/ |
|
|
| Back to top |
|
|
|
| Nathan Sanders... |
| Posted: Sun Nov 08, 2009 12:38 am |
|
|
|
Guest
|
In article <hd3eqj$eg7$02$2 at (no spam) news.t-online.com>,
Joachim Pense <snob at (no spam) pense-mainz.eu> wrote:
[quote]Nathan Sanders (in sci.lang):
In article <87fx8rgg7m.fsf at (no spam) gmail.com>,
LEE Sau Dan <danlee at (no spam) informatik.uni-freiburg.de> wrote:
"Nathan" == Nathan Sanders <nsanders at (no spam) williams.edu> writes:
No. You seem to consider mathematical notation as a
prescriptivist considers English.
Nathan> Of course. Mathematical notation, unlike language, was
Nathan> consciously and intentionally designed.
Designed by whom? God?
By humans, obviously.
If you realized that it is designed by human beings, would you still
suppose that it is flawless?
I don't "suppose" it; I know it.
Are you really unaware that mathematical logic defines the notion of
well-formed formula?
Can humans design flawed notation systems? Of course. Those are
broken systems that good mathematicians would not use, because such
systems result in exactly the kind of ambiguity and imprecision that
mathematical notation is ordinarily designed to avoid. How terrible
it would be to have a mathematical notation with no defined order of
operations or grouping!
Why is ambiguity in language more acceptable?
[/quote]
Because we have no control over the fundamental nature of language.
As rational people, we accept the properties of language, because
those properties are given to us as facts.
(Okay, not all of us accept those properties. Prescriptivists do
often try to change some of them, but it's well-known how unsuccessful
those attempts generally are, especially in spoken language.)
But formal languages are invented abstract objects (like Magdalenian
and Zamenhof's Esperanto), which means they have no intrinsic need to
adhere to the human brain's linguistic quirks. Free from these
limitations, we can make formal languages as different from human
language as our imaginations will allow. This includes allowing them
to lack ambiguity.
Nathan
--
Nathan Sanders
Linguistics Program
Williams College
http://wso.williams.edu/~nsanders/ |
|
|
| Back to top |
|
|
|
| Joachim Pense... |
| Posted: Sun Nov 08, 2009 2:38 am |
|
|
|
Guest
|
António Marques (in sci.lang):
[quote]On Nov 7, 12:46Â pm, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
António Marques (in sci.lang):
On 7 Nov, 06:05, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
António Marques (in sci.lang):
On Nov 6, 9:19Â pm, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
António Marques (in sci.lang):
On 6 Nov, 18:44, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
Why do you think higher level
programming languages exist? They have been invented because it
is easier to write complex programs in them than it is if you
write machine code that does the equivalent job. Some
programming languages are flexible in the code they accept
precisely to allow for pragmatics in the code.
I must be missing something. What exactly (in programming
languages) are you talking about?
Pragmatics. I had given an example before.
Must have missed it. Be so kind as to elaborate, give another
example, etc.
The following lines of Perl all do the same thing but are
pragmatically different:
if ($a != 1) {$b = 17}
unless ($a == 1) {$b = 17}
$b = 17 unless $a == 1
$a == 1 or $b = 17
$a != 1 and $b = 17
I see what you mean. Sorry I hadn't spotted it earlier and made you
repeat it.
However, I don't think these are examples of pragmatics. They are
structurally different ways of getting the exact same result; their
intent, as you mention, is to better model what's on the programmer's
mind - say, make the language more semantic programmer-wise. In what
way do you consider it to be an example of pragmatics?
In my view these are examples of topicalizing a sentence differently.
Isn't topicalization part of pragmatics?
I think "better model what's on the programmer's mind" is more a
pragmatic than a semantic concept.
But pragmatics has to do with how the receiver determines the
speaker's intent. While in your examples the computer always has only
one possible interpretation.
[/quote]
The receiver we are talking about here is not the computer but the human
reader of a program.
[quote]Unless you are discussing programming
languages bot as a medium for programmer <> computer interaction, but
as a way to convey information between programmers.
[/quote]
Of course I do.
[quote]In the latter
case, it's only natural that some properties of human language trickle
down into it - but still, in my view, they get there not because of
the need for programmer <> programmer interaction, but as an
expressiveness convenience for the original programmer, regardless of
whether the code will ever again be analyzed by a human or not.
[/quote]
The pragmatical features of human language (e.g. topicalization) also apply
to inner monologues and diary texts that are not supposed to be read by
anybody else.
Joachim |
|
|
| Back to top |
|
|
|
| Joachim Pense... |
| Posted: Sun Nov 08, 2009 3:07 am |
|
|
|
Guest
|
Nathan Sanders (in sci.lang):
[quote]In article <hd3q1s$n85$01$1 at (no spam) news.t-online.com>,
Joachim Pense <snob at (no spam) pense-mainz.eu> wrote:
António Marques (in sci.lang):
On 7 Nov, 06:05, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
António Marques (in sci.lang):
On Nov 6, 9:19Â pm, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
António Marques (in sci.lang):
On 6 Nov, 18:44, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
Why do you think higher level
programming languages exist? They have been invented because it
is easier to write complex programs in them than it is if you
write machine code that does the equivalent job. Some
programming languages are flexible in the code they accept
precisely to allow for pragmatics in the code.
I must be missing something. What exactly (in programming
languages) are you talking about?
Pragmatics. I had given an example before.
Must have missed it. Be so kind as to elaborate, give another
example, etc.
The following lines of Perl all do the same thing but are
pragmatically different:
if ($a != 1) {$b = 17}
unless ($a == 1) {$b = 17}
$b = 17 unless $a == 1
$a == 1 or $b = 17
$a != 1 and $b = 17
I see what you mean. Sorry I hadn't spotted it earlier and made you
repeat it.
However, I don't think these are examples of pragmatics. They are
structurally different ways of getting the exact same result; their
intent, as you mention, is to better model what's on the programmer's
mind - say, make the language more semantic programmer-wise. In what
way do you consider it to be an example of pragmatics?
In my view these are examples of topicalizing a sentence differently.
Isn't topicalization part of pragmatics?
It is, but is it part of programming languages? What observable,
functional role does it play? If the compiler cares, then it's not
pragmatics; if the compiler doesn't care, then why should we?
[/quote]
There are readers of the program; most important, the programmer herself who
has to understand the code one hour later, And then auditors who check the
code, other programmers who maintain the code.
And, as Antonio wrote, it is there to give the programmer the expressive
power to formulate his ideas and concepts conveniently. Even more so, the
concepts are in the source program, but they are lost in the process of
compilation.
(Sometimes there isn't even a computer involved. Sau-Dan mentioned
Pseudocode; I add that the programming language Algol was designed at the
end of the 50 as an "Algorithmic Language", with the intention to allow
mathematicians to write down numerical algorithms. It even didn't have
input-output statements.)
[quote]But even if something like pragmatics plays an identifiable role in
computer programming, is it really an inherent part of the computer
language itself, or just something copied over from the programmer's
own native language? That is, do all typical programmers writing in
the same computer language for the same compilers have identical
pragmatic behavior, or does their pragmatic behavior vary based on
their native language? (Or worse, does it vary from individual to
individual in unpredictable ways? If so, that's not pragmatics.)
[/quote]
Interesting question. I never observed anything like this that could give me
a hint how to answer it. The _programming language_ one was trained in
first sometimes show in the way someone programs in other programming
languages.
[quote]
I think "better model what's on the programmer's mind" is more a
pragmatic than a semantic concept.
I remain unconvinced that, in the context of the properties of
computer languages, "what's on the programmer's mind" is a concept
worth making a model of, or even possible to make a useful model of.
Is "what's on the programmer's mind" really consistent and predictable
enough to even be meaningfully modeled?
[/quote]
That's what the programmer does when he designs his program.
Joachim |
|
|
| Back to top |
|
|
|
| Joachim Pense... |
| Posted: Sun Nov 08, 2009 3:18 am |
|
|
|
Guest
|
Nathan Sanders (in sci.lang):
[quote]In article <hd3eqj$eg7$02$2 at (no spam) news.t-online.com>,
Joachim Pense <snob at (no spam) pense-mainz.eu> wrote:
Nathan Sanders (in sci.lang):
In article <87fx8rgg7m.fsf at (no spam) gmail.com>,
LEE Sau Dan <danlee at (no spam) informatik.uni-freiburg.de> wrote:
"Nathan" == Nathan Sanders <nsanders at (no spam) williams.edu> writes:
No. You seem to consider mathematical notation as a
prescriptivist considers English.
Nathan> Of course. Mathematical notation, unlike language, was
Nathan> consciously and intentionally designed.
Designed by whom? God?
By humans, obviously.
If you realized that it is designed by human beings, would you
still suppose that it is flawless?
I don't "suppose" it; I know it.
Are you really unaware that mathematical logic defines the notion of
well-formed formula?
Can humans design flawed notation systems? Of course. Those are
broken systems that good mathematicians would not use, because such
systems result in exactly the kind of ambiguity and imprecision that
mathematical notation is ordinarily designed to avoid. How terrible
it would be to have a mathematical notation with no defined order of
operations or grouping!
Why is ambiguity in language more acceptable?
Because we have no control over the fundamental nature of language.
As rational people, we accept the properties of language, because
those properties are given to us as facts.
[/quote]
You are saying that we accept that a language allows ambiguity as a fact of
life, but does not mean we accept instances of ambiguity. We try to avoid
them when we use language, because we don't want to be misunderstood (or
maybe if we want to cheat we exploit ambiguities of language as in the
example of the man who paid for "intercourse with a young lady" and later
unsuccessfully complained that the lady only wanted to chat with him).
Of course a well designed notation system or programming language offers
almost no possibilities for ambiguity, but if they are present, they have
to be avoided.
Joachim |
|
|
| Back to top |
|
|
|
| Joachim Pense... |
| Posted: Sun Nov 08, 2009 3:25 am |
|
|
|
Guest
|
DKleinecke (in sci.lang):
[quote]On Nov 7, 1:41Â am, Joachim Pense <s... at (no spam) pense-mainz.eu> wrote:
Nathan Sanders (in sci.lang):
In article <877hu2g7f0.... at (no spam) informatik.uni-freiburg.de>,
LEE Sau Dan <dan... at (no spam) informatik.uni-freiburg.de> wrote:
"Nathan" == Nathan Sanders <nsand... at (no spam) williams.edu> writes:
Maybe, you could try using the postfix notation instead to see if
that helps stopping them from thinking so. Â When people see
something unfamiliar (postfix notation in this case), they may
refrain from generalizing their past knowledge into the new
field.
Nathan> Why would I introduce a completely different notation from
Nathan> that that is (a) most common and (b) used in the textbook?
Not a  completely different one.  But  you may need  to use
non-standard things for:
1) new concepts;
Actually, we do cover concepts not formalized in the textbook
(pragmatics), and I do introduce new formal notation for those
concepts. Â However, this comes near the end of the semester, after
they're already long-accustomed to the standard notation in the
textbook (infixed binary logical connectives and prefixed quantifiers,
negation, and predicates), so I'm not going to switch things up on
them in the last few weeks. Â Especially not with postfix notation!
2) convenience (to introduce short-hands, such as Einstein's notation)
The point of the formal definitions is to have precision. Â Convenience
is nice, but is not the primary goal (or even a goal; it doesn't
really matter at all how "convenient" the logical system is).
I hated the kind of mathematics that depends on the correctness of long
algebraic calculations. If I have to believe a fact that is only founded
on manipulation of formulas, I don't get the insight why it is true, and
so this is error-prone. I preferred arriving at results by insights and
using formulas only as a second step putting those insights on paper,
hopefully invoking the insight in the reader. Normally, prose works
better for that purpose.
Unless you have an insight into the distribution of prime numbers I
don't see how you come to believe any theorem about the distribution
of prime numbers. If you have an insight into the distribution of
prime numbers please share it with us.
[/quote]
I leave prime numbers to those who care about them.
Joachim |
|
|
| Back to top |
|
|
|
| LEE Sau Dan... |
| Posted: Sun Nov 08, 2009 6:14 am |
|
|
|
Guest
|
[quote]"António" == António Marques <entonio at (no spam) gmail.com> writes:
[/quote]
António> The outcome is predictable, since the outcome is known to
António> be unordered. So, 'A B C D' and 'A C D B' are the 'same'
António> outcome. Only the properties expected (and here order is
António> not one of them) are relevant. (I actually should know; I
António> routinely spend time reading database rows into objects and
António> putting those into hashtables - which, again, usually have
António> no ordering - and it's all the same whether A or B comes
António> first, either in the original row set or in the resulting
António> hashtable. Whenever order matters, the query must be
António> different and the hashtable must present itself sorted by
António> order of insertion.)
A hashtable that maintains insertion order? Is that still a hashtable?
--
Lee Sau Dan æŽå®ˆæ•¦ ~{ at (no spam) nJX6X~}
E-mail: danlee at (no spam) informatik.uni-freiburg.de
Home page: http://www.informatik.uni-freiburg.de/~danlee |
|
|
| Back to top |
|
|
|
| LEE Sau Dan... |
| Posted: Sun Nov 08, 2009 6:25 am |
|
|
|
Guest
|
[quote]"Joachim" == Joachim Pense <snob at (no spam) pense-mainz.eu> writes:
But even if something like pragmatics plays an identifiable role
in computer programming, is it really an inherent part of the
computer language itself, or just something copied over from the
programmer's own native language? That is, do all typical
programmers writing in the same computer language for the same
compilers have identical pragmatic behavior, or does their
pragmatic behavior vary based on their native language? (Or
worse, does it vary from individual to individual in
unpredictable ways? If so, that's not pragmatics.)
[/quote]
Joachim> Interesting question. I never observed anything like this
Joachim> that could give me a hint how to answer it. The
Joachim> _programming language_ one was trained in first sometimes
Joachim> show in the way someone programs in other programming
Joachim> languages.
Yes. This is especially observable when the "other" programming
language uses a different paradigm. Very often, the programmer is
unable to shift to the new paradigm, and would try to make the new
things look like nails to suit his old hammer. e.g. I've seen
Postscript code (a stack-oriented language) written in a very C-like
style. (C is imperative, (local) variable-based.)
[quote]I think "better model what's on the programmer's mind" is more a
pragmatic than a semantic concept.
I remain unconvinced that, in the context of the properties of
computer languages, "what's on the programmer's mind" is a
concept worth making a model of, or even possible to make a
useful model of. Is "what's on the programmer's mind" really
consistent and predictable enough to even be meaningfully
modeled?
[/quote]
Joachim> That's what the programmer does when he designs his
Joachim> program.
Yes. A person who has no experience with a stack-oriented language
(e.g. Forth, Postscript), for instance, would write Postscript code in a
very C-like style. He would not natively use the stack (or even think
about it). He would try to immitate local variables by using a new
dictionary. etc.
See this thread for an example:
http://groups.google.co.uk/group/comp.lang.postscript/browse_thread/thread/a26b9bbebdc678f7
--
Lee Sau Dan §õ¦u´° ~{ at (no spam) nJX6X~}
E-mail: danlee at (no spam) informatik.uni-freiburg.de
Home page: http://www.informatik.uni-freiburg.de/~danlee |
|
|
| Back to top |
|
|
|
|
|
All times are GMT - 5 Hours
The time now is Tue Dec 08, 2009 10:42 am
|
|