New Science Discovery: Perl Idiots Remain Idiots After A Decade!New

X

Xah Lee

New Science Discovery: Perl Idiots Remain Idiots After A Decade!

A excerpt from the new book 〈Modern Perl〉, just published, chapter 4
on “Operatorsâ€. Quote:

«The associativity of an operator governs whether it evaluates from
left to right or right to left. Addition is left associative, such
that 2 + 3 + 4 evaluates 2 + 3 first, then adds 4 to the result.
Exponentiation is right associative, such that 2 ** 3 ** 4 evaluates 3
** 4 first, then raises 2 to the 81st power. »

LOL. Looks like the perl folks haven't changed. Fundamentals of
serious math got botched so badly.

Let me explain the idiocy.

It says “The associativity of an operator governs whether it evaluates
from left to right or right to left.â€. Ok, so let's say we have 2
operators: a white triangle â–³ and a black triangle â–². Now, by the
perl's teaching above, let's suppose the white triangle is “right
associative†and the black triangle is “left associativeâ€. Now, look
at this:

3 â–³ 6 â–² 5

seems like the white and black triangles are going to draw a pistol
and fight for the chick 6 there. LOL.

Now, let me tell you what operator precedence is. First of all, let's
limit ourselfs to discuss operators that are so-called binary
operators, which, in our context, basically means single symbol
operator that takes it's left and right side as operands. Now, each
symbol have a “precedenceâ€, or in other words, the set of operators
has a order. (one easy way to think of this is that, suppose you have
n symbols, then you give each a number, from 1 to n, as their order)
So, when 2 symbols are placed side by side such as 「3 â–³ 6 â–² 5ã€, the
symbol with higher precedence wins. Another easy way to think of this
is that each operator has a stickiness level. The higher its level, it
more sticky it is.

the problem with the perl explanations is that it's one misleading
confusion ball. It isn't about “left/right associativityâ€. It isn't
about “evaluates from left to right or right to leftâ€. Worse, the word
“associativity†is a math term that describe a property of algebra
that has nothing to do with operator precedence, yet is easily
confused with because it is a property about order of evaluation. (for
example, the addition function is associative, meaning: 「(3+6)+5 =
3+(6+5)ã€.)

compare it with this:

〈Perl & Python: Complex Numbers〉
http://xahlee.org/perl-python/complex_numbers.html

and for a good understanding of functions and operators, see:

〈What's Function, What's Operator?〉
http://xahlee.org/math/function_and_operators.html
 
C

Chiron

On Wed, 29 Feb 2012 00:09:16 -0800, Xah Lee wrote:

Personally, I think this whole issue of precedence in a programming
language is over-rated. It seems to me that grouping of any non-trivial
set of calculations should be done so as to remove any possible confusion
as to intent. It is one more obstacle to accidental errors in logic,
where you intend one thing, possibly overlook precedence, and get a
strange result.

Sure, mathematically it *should* go a particular way, and any programming
language *should* follow that. Still... they don't, and since they don't
it makes more sense to be really obvious what you meant to do.

As someone pointed out, a programming language is for humans; computers
don't need them. That being the case, it makes sense to keep things as
clear as possible.
 
R

Rainer Weikusat

Xah Lee said:
A excerpt from the new book 〈Modern Perl〉, just published, chapter 4
on “Operatorsâ€. Quote:

«The associativity of an operator governs whether it evaluates from
left to right or right to left. Addition is left associative, such
that 2 + 3 + 4 evaluates 2 + 3 first, then adds 4 to the result.
Exponentiation is right associative, such that 2 ** 3 ** 4 evaluates 3
** 4 first, then raises 2 to the 81st power. »

LOL. Looks like the perl folks haven't changed. Fundamentals of
serious math got botched so badly.

Let me explain the idiocy.

It says “The associativity of an operator governs whether it evaluates
from left to right or right to left.â€. Ok, so let's say we have 2
operators: a white triangle â–³ and a black triangle â–². Now, by the
perl's teaching above, let's suppose the white triangle is “right
associative†and the black triangle is “left associativeâ€. Now, look
at this:

3 â–³ 6 â–² 5

seems like the white and black triangles are going to draw a pistol
and fight for the chick 6 there. LOL.

As the perlop manpage would have told you,

Operator associativity defines what happens if a sequence of the same
operators is used one after another

Since this is not the case in your example, it doesn't seem to be
applicable here. Also, the Perl I'm aware doesn't have 'white
triangle' and 'black triangle' operators and it also doesn't have
operators of equal precedence and different associativity. It can't,
actually, since there would be no way to evaluate an expression like
the mock one you invented above. Lastly, that something happens to be
in one way or another way in the completely arbitrary set of rules and
conventions commonly referred to as 'mathematics' (an essentially
outdated write-only programming language dating back to the times
when humans had to perform computations themselves) doesn't mean it is
of any relevance anywhere else just because of this, no matter how
dear it might be to lots of people.
 
C

Chiron

[...] 'mathematics' (an essentially
outdated write-only programming language dating back to the times when
humans had to perform computations themselves) [...]

Theoretical Computer Science is a branch of mathematics. Are you saying
it is outdated?

Kiuhnm

Neither mathematics nor computer science is outdated. Such an assertion
is without merit.

Mathematics is not exclusively - nor even primarily - concerned with
computations.
 
N

namekuseijin

New Science Discovery: Perl Idiots Remain Idiots After A Decade!

A excerpt from the new book 〈Modern Perl〉, just published, chapter 4
on “Operatorsâ€. Quote:

«The associativity of an operator governs whether it evaluates from
left to right or right to left. Addition is left associative, such
that 2 + 3 + 4 evaluates 2 + 3 first, then adds 4 to the result.
Exponentiation is right associative, such that 2 ** 3 ** 4 evaluates 3
** 4 first, then raises 2 to the 81st power. »

LOL. Looks like the perl folks haven't changed. Fundamentals of
serious math got botched so badly.

Let me explain the idiocy.

It says “The associativity of an operator governs whether it evaluates
from left to right or right to left.â€. Ok, so let's say we have 2
operators: a white triangle â–³ and a black triangle â–². Now, by the
perl's teaching above, let's suppose the white triangle is “right
associative†and the black triangle is “left associativeâ€. Now, look
at this:

3 â–³ 6 â–² 5

seems like the white and black triangles are going to draw a pistol
and fight for the chick 6 there. LOL.

Now, let me tell you what operator precedence is. First of all, let's
limit ourselfs to discuss operators that are so-called binary
operators, which, in our context, basically means single symbol
operator that takes it's left and right side as operands. Now, each
symbol have a “precedenceâ€, or in other words, the set ofoperators
has a order. (one easy way to think of this is that, suppose you have
n symbols, then you give each a number, from 1 to n, as their order)
So, when 2 symbols are placed side by side such as 「3 â–³ 6â–² 5ã€, the
symbol with higher precedence wins. Another easy way to think of this
is that each operator has a stickiness level. The higher its level, it
more sticky it is.

the problem with the perl explanations is that it's one misleading
confusion ball. It isn't about “left/right associativityâ€.. It isn't
about “evaluates from left to right or right to leftâ€. Worse, the word
“associativity†is a math term that describe a property of algebra
that has nothing to do with operator precedence, yet is easily
confused with because it is a property about order of evaluation. (for
example, the addition function is associative, meaning: 「(3+6)+5 =
3+(6+5)ã€.)

compare it with this:

〈Perl & Python: Complex Numbers〉http://xahlee.org/perl-python/complex_numbers.html

and for a good understanding of functions and operators, see:

〈What's Function, What's Operator?〉http://xahlee.org/math/function_and_operators.html

associativity of operators mean little in the Lisp world obviously, so
why was this posted here? Sorry, perl, python and emacs folks...

BTW, it's the same in javascript: it is so such that 2 + 3 + "4" is
"54" and "2" + 3 + 4 is "234". Blame weak typing and + overloading,
though it may be a blessing.
 
X

Xah Lee

i missed a point in my original post. That is, when the same operator
are adjacent. e.g. 「3 â–² 6 â–² 5ã€.

This is pointed out by Kiuhnm 〔kiuhnm03.4t.yahoo.it〕 and Tim Bradshaw.
Thanks.

though, i disagree the way they expressed it, or any sense this is
different from math.

to clarify, amend my original post, here's what's needed for binary
operator precedence:

â‘  the symbols are ordered. (e.g. given a unique integer)

â‘¡ each symbol is has either one of left-side stickness or right-side
stickness spec. (needed when adjacent symbols are the same.)

About the lisp case mentioned by Tim, e.g. in「(f a b c)ã€, whether it
means 「(f (f a b) c)〠or 「(f a (f b c))〠. It is not directly relevant
to the context of my original post, because it isn't about to
operators. It's about function argument eval order. Good point,
nevertheless.

the perl doc, is still misleading, terribly bad written. Becha ass!

Xah
 
C

Chiron

No. Mathematically it should go the way that it is defined to go. There
is nothing in Mathematics that either requires or prohibits infix
notation in programming languages, or even in Mathematical notation.
Yes. That (the mathematically defined way) is a particular way, is it
not?
Often infix notation with well thought out precedence is the clearest
way to go. RPN and the like have their place, but often are difficult
for real people to read.

However, I wasn't specifically referring to infix/postfix/prefix or
anything of that nature. I wasn't limiting my comment to lisp notation
in particular, since what I said applies to any language. I was
referring to the placement of parentheses (or other groupings) to
indicate to *humans* what the intended sequence of events was. The
problem with precedence is that it is not always clear how it will go.
Different languages have different rules, some of which depart from the
rules in mathematics. Some implementations of languages are buggy in
this regard.

Mathematically, and in any language with which I am familiar, the
sequence: 2 + 6 / 3 will yield 4. It is unnecessary, but harmless, to
write this as 2 + (6 / 3). A naive reader (or just a tired or hurried
one) might come up with 8 / 3 if there aren't any parentheses.

Whenever there is *any* possibility of ambiguity, I see no reason not to
clarify. Back in the days when the way you wrote your code affected how
it was compiled, it made sense to rely heavily on language-specific
features, thus saving a few bytes. With gigabyte memories, gigahertz
clock speeds, and optimizing compilers, the pressure to try to optimize
by hand is gone. A few extra parentheses, or even breaking down a
complex sequence of events into discrete, simpler ones, is no longer a
costly luxury. A few extra variables, if they help clarity, aren't going
to hurt anything. Let the machine do the grunt work. Pamper your
readers (which in a few weeks or months might be you) and show exactly
what you had in mind. That's all I'm saying.
 
C

Chiron

ROTF,LMAO! You obviously don't have a clue as to what Mathematics means.
Free hint: it doesn't mean Arithmetic. You're as bigoted as Xah Lee,


Hmm... maybe, instead of just ridiculing him, you could explain where he
is mistaken. Of course, doing that is a *LOT* harder than just calling
him a bigot.

BTW, I happen to agree with you insofar as this poster not understanding
the nature of mathematics. His comment reminds me of the article,
"Transgressing the Boundaries: Towards a Transformative Hermeneutics of
Quantum Gravity" (http://www.physics.nyu.edu/sokal/transgress_v2/
transgress_v2_singlefile.html). Also known as the "Sokal Hoax."
 
R

Rainer Weikusat

Shmuel (Seymour J.) Metz said:
ROTF,LMAO! You obviously don't have a clue as to what Mathematics
means. Free hint: it doesn't mean Arithmetic.

You obviously don't have any sense of humour. But don't let this
trouble you.
 
C

Chiron

On Wed, 29 Feb 2012 00:09:16 -0800, Xah Lee wrote:

Xah, you won't grow even an inch taller by cutting others down.
 
X

Xah Lee

They did not make up the terminology, if that is what you are saying.
The concepts of left and right associativity are well-known and accepted
in TCS (Theoretical CS).

Aho, Sethi and Ullman explain it this way in "Compilers: Principles,
Techniques and Tools":
"We say that the operator + associates to the left because an operand
with plus signs on both sides of it is taken by the operator to its
left. [...]"
And they also show parse trees similar to the ones I wrote above.

how do they explain when 2 operators are adjacent e.g. 「3 △6 ▲ 5 �

do you happen to know some site that shows the relevant page i can
have a look?

thanks.

Xah

i missed a point in my original post. That is, when the same operator
are adjacent. e.g. 「3 â–² 6 â–² 5ã€.
This is pointed out by Kiuhnm 〔kiuhnm03.4t.yahoo.it〕 and Tim Bradshaw.
Thanks.
though, i disagree the way they expressed it, or any sense this is
different from math.

They did not make up the terminology, if that is what you are saying.
The concepts of left and right associativity are well-known and accepted
in TCS (Theoretical CS).

If you change the terminology, no one will understand you unless you
provide your definitions every time (and then they may not accept them).

Another way of saying that an operator is left-associative is that its
parse tree is a left-tree, i.e. a complete tree where each right child
is a leaf.
For instance, (use a monospaced font)
   1 + 2 + 3 + 4
gives you this left-tree:
       +
     +   4
   +   3
  1 2
while 1**2**3**4
gives you this right-tree:
   **
1    **
    2    **
        3  4

Aho, Sethi and Ullman explain it this way in "Compilers: Principles,
Techniques and Tools":
"We say that the operator + associates to the left because an operand
with plus signs on both sides of it is taken by the operator to its
left. [...]"
And they also show parse trees similar to the ones I wrote above.

Kiuhnm
 
C

Chiron

I'm treating him as he treats others.
OK.

A brilliant piece of work. I greatly enjoyed it and the reaction to its
disclosure.

What always gets me is how so many people criticized Sokal for doing it,
instead of soundly condemning the editor for not bothering to verify what
Sokal said. It's like the kid points out that the emperor has no
clothes, so they shoot the kid. Of course, in real life, that's exactly
what would happen, so I guess I shouldn't be too surprised...
 
C

Chiron

Google for Omerta. It's common for whistle blowers to be chastised or
even persecuted. I agree that the criticism of Prof Sokal was
outrageous, but it was also predictable.

Yeah, omerta... I'm familiar with it. Talk and you're dead, and they put
a canary in your mouth (well, some folks do, anyway).

But of course you're right - it's a milder form of omerta. It's just so
misguided. Kill the messenger. Imprison the whistle-blower.
 
A

Albert van der Horst

New Science Discovery: Perl Idiots Remain Idiots After A Decade!

A excerpt from the new book =E3=80=88Modern Perl=E3=80=89, just published, =
chapter 4
on =E2=80=9COperators=E2=80=9D. Quote:

=C2=ABThe associativity of an operator governs whether it evaluates from
left to right or right to left. Addition is left associative, such
that 2 + 3 + 4 evaluates 2 + 3 first, then adds 4 to the result.
Exponentiation is right associative, such that 2 ** 3 ** 4 evaluates 3
** 4 first, then raises 2 to the 81st power. =C2=BB

LOL. Looks like the perl folks haven't changed. Fundamentals of
serious math got botched so badly.

You're confused.

Associativity of operators is defined in mathematics.
(The same concept may be used in programming).
"left-associativity" and "right-associativity" are computer languages
concept and their definitions are not from mathematics.

Interestingly in mathematics associative means that it doesn't matter
whether you use (a.b).c or a.(b.c).
Using xxx-associativity to indicate that it *does* matter is
a bit perverse, but the Perl people are not to blame if they use
a term in their usual sense.

Groetjes Albert
 
R

Raymond Wiker

Shmuel (Seymour J.) Metz said:
No, s/h/it is just an acephalic troll with delusions of adequacy.

Another way to put it is to say that Xah is a legend in his
own mind.
 
A

Albert van der Horst

You may see it this way:
Def1. An operator +:SxS->S is left-associative iff
a+b+c = (a+b)+c for all a,b,c in S.
Def2. An operator +:SxS->S is right-associative iff
a+b+c = a+(b+c) for all a,b,c in S.
Def3. An operator +:SxS->S is associative iff it is both left and
right-associative.

I know, but what the mathematicians do make so much more sense:
(a+b)+c = a+(b+c) definition of associative.
Henceforth we may leave out the brackets.

Don't leave out the brackets if the operators if the operators is
not associative.

P.S. There is no need for the operators to be SxS->S.
For example a b c may be m by n, n by l, l by k matrices respectively.

Groetjes Albert
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

No members online now.

Forum statistics

Threads
473,743
Messages
2,569,478
Members
44,898
Latest member
BlairH7607

Latest Threads

Top