VERY simple question about "?"

ruby-talk
1

I absolutely love Ruby, but...I've always found the subject of Ruby operators to be a small nightmare. First, it's hard to find a complete table - the best I've found says gleefully at its bottom "and there are even more". Wonderful. So many tools, so little time...

Then there's the matter of some symbols being operators, others being methods, and some (my impression only) both, and yet others being a number of things, depending upon context. When someone says that Ruby is harder than Python, I suppose this might be what they mean. makes sense to me.

So, just today I learn that this is possible:

irb(main):001:0> ?.
=> 46
irb(main):002:0> ?a
=> 97
irb(main):003:0> 97.chr
=> "a"
irb(main):004:0>

So...what is this "?" thing? I've looked in four references, and Googled until I'm dizzy, and find no answer. Operator? Method? (Stupid distinction???)

I know I'll certainly be using this, but it'd be nice to know what I'm doing.

All comments welcome.

t.

···

--

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd, MS MA, LMHC - Private practice Psychotherapist
Bellingham, Washington, U.S.A: (360) 920-1226
<< tc@tomcloyd.com >> (email)
<< TomCloyd.com >> (website) << sleightmind.wordpress.com >> (mental health weblog)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

2

It's simply a language construct, like the ' that delimitates a string
or the : that starts a symbol.

If you have the pickaxe book, it's documented in the "Integer and
Floating-Point Numbers" section (page 304 and 305 of my PDF edition).

Fred

···

Le 4 janvier 2009 à 12:58, Tom Cloyd a écrit :

So...what is this "?" thing? I've looked in four references, and Googled
until I'm dizzy, and find no answer. Operator? Method? (Stupid
distinction???)

--
Go not to Usenet for counsel, for they will say both 'No' and 'Yes' and
'Try another newsgroup'.

3

And if you don't have the book, you can read that section[1] in the online
version.

Regards,
Yaser

[1]: http://whytheluckystiff.net/ruby/pickaxe/html/language.html#UC

···

On Sun, Jan 4, 2009 at 3:34 PM, F. Senault <fred@lacave.net> wrote:

If you have the pickaxe book, it's documented in the "Integer and
Floating-Point Numbers" section (page 304 and 305 of my PDF edition).

4

Yaser Sulaiman wrote:

If you have the pickaxe book, it's documented in the "Integer and
Floating-Point Numbers" section (page 304 and 305 of my PDF edition).

And if you don't have the book, you can read that section[1] in the online
version.

Regards,
Yaser

[1]: http://whytheluckystiff.net/ruby/pickaxe/html/language.html#UC

Thanks, guys.

I guess what confused me is that I bring to Ruby the notion that if a symbol causes something to happen, as "?" clearly does in my example (it produces the integer code for a character), its' an operator or method. I'm not unfamiliar with linguistics. I don't see the equivalence between a string delimiter, or a character that signals the beginning of a symbol, and a symbol that is actually productive of something. Makes no sense to me at all, in fact. Maybe I'm failing to grasp some transformative concept which changes how things work in the Ruby world.

Seeing operators as methods makes some real sense, on the other hand. "=" produces something out of what it's given. It doesn't just change meaning, which is all that happens with quotes or ":". That's the distinction I'm seeing, and it seems valid.

I'll keep thinking about it.

Quite apart from this is the strange matter of how difficult it is to find a simple, complete table of Ruby operators. I've basically given up on that.

Thanks for the Pickaxe reference - I found it on PDF p. 332 of my PDF version of the 2nd ed. I'd never seen that page.

t.

···

On Sun, Jan 4, 2009 at 3:34 PM, F. Senault <fred@lacave.net> wrote:

--

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd, MS MA, LMHC - Private practice Psychotherapist
Bellingham, Washington, U.S.A: (360) 920-1226
<< tc@tomcloyd.com >> (email)
<< TomCloyd.com >> (website) << sleightmind.wordpress.com >> (mental health weblog)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

5

I guess what confused me is that I bring to Ruby the notion that if a symbol causes something to happen, as "?" clearly does in my example (it produces the integer code for a character),

In this case nothing "happens" (see below). It is just another syntactical way to express an integer constant.

its' an operator or method. I'm not unfamiliar with linguistics. I don't see the equivalence between a string delimiter, or a character that signals the beginning of a symbol, and a symbol that is actually productive of something. Makes no sense to me at all, in fact. Maybe I'm failing to grasp some transformative concept which changes how things work in the Ruby world.

Seeing operators as methods makes some real sense, on the other hand. "=" produces something out of what it's given. It doesn't just change meaning, which is all that happens with quotes or ":". That's the distinction I'm seeing, and it seems valid.

I am not sure what you mean by "change meaning". Can you elaborate?

Btw, note that there are two types of "=":

1. variable assignment as in "a = 123",

2. method call that mimics variable assignment as in "a.foo = 456".

Both evaluate always to the right hand side - but this is about the only commonality. Semantics of type 1 is fixed while you are free to implement type 2 in any way you like, e.g.

irb(main):005:0> a = Object.new
=> #<Object:0x7ff79980>
irb(main):006:0> def a.foo=(x) puts "bah!" end
=> nil
irb(main):007:0> a.foo = 456
bah!
=> 456
irb(main):008:0>

I'll keep thinking about it.

Basically "?a" is just another textual representation for the integer 97 which happens to be the ASCII code of character "a". It differs from string delimiters because the expression "foo" is actually a String constructor - it will create a new String object whenever it is executed:

irb(main):003:0> 3.times { x = "foo"; puts x, x.object_id }
foo
1073491940
foo
1073492040
foo
1073491880
=> 3
irb(main):004:0>

Quite apart from this is the strange matter of how difficult it is to find a simple, complete table of Ruby operators. I've basically given up on that.

You can find it in the Pickaxe - a good book to have anyway.

Thanks for the Pickaxe reference - I found it on PDF p. 332 of my PDF version of the 2nd ed. I'd never seen that page.

As far as I know there is no legal electronic version of the second edition so you better double check whether you are allowed to use this. You could as well buy the book to somewhat compensate for this.

Kind regards

  robert

···

On 04.01.2009 14:48, Tom Cloyd wrote:

6

There definitely is an option to select a PDF version for purchase.

···

On Sun, Jan 4, 2009 at 9:48 AM, Robert Klemme <shortcutter@googlemail.com> wrote:

On 04.01.2009 14:48, Tom Cloyd wrote:

I guess what confused me is that I bring to Ruby the notion that if a
symbol causes something to happen, as "?" clearly does in my example (it
produces the integer code for a character),

In this case nothing "happens" (see below). It is just another syntactical
way to express an integer constant.

its' an operator or method. I'm not unfamiliar with linguistics. I don't
see the equivalence between a string delimiter, or a character that signals
the beginning of a symbol, and a symbol that is actually productive of
something. Makes no sense to me at all, in fact. Maybe I'm failing to grasp
some transformative concept which changes how things work in the Ruby world.

Seeing operators as methods makes some real sense, on the other hand. "="
produces something out of what it's given. It doesn't just change meaning,
which is all that happens with quotes or ":". That's the distinction I'm
seeing, and it seems valid.

I am not sure what you mean by "change meaning". Can you elaborate?

Btw, note that there are two types of "=":

1. variable assignment as in "a = 123",

2. method call that mimics variable assignment as in "a.foo = 456".

Both evaluate always to the right hand side - but this is about the only
commonality. Semantics of type 1 is fixed while you are free to implement
type 2 in any way you like, e.g.

irb(main):005:0> a = Object.new
=> #<Object:0x7ff79980>
irb(main):006:0> def a.foo=(x) puts "bah!" end
=> nil
irb(main):007:0> a.foo = 456
bah!
=> 456
irb(main):008:0>

I'll keep thinking about it.

Basically "?a" is just another textual representation for the integer 97
which happens to be the ASCII code of character "a". It differs from string
delimiters because the expression "foo" is actually a String constructor -
it will create a new String object whenever it is executed:

irb(main):003:0> 3.times { x = "foo"; puts x, x.object_id }
foo
1073491940
foo
1073492040
foo
1073491880
=> 3
irb(main):004:0>

Quite apart from this is the strange matter of how difficult it is to find
a simple, complete table of Ruby operators. I've basically given up on that.

You can find it in the Pickaxe - a good book to have anyway.

Thanks for the Pickaxe reference - I found it on PDF p. 332 of my PDF
version of the 2nd ed. I'd never seen that page.

As far as I know there is no legal electronic version of the second edition
so you better double check whether you are allowed to use this. You could
as well buy the book to somewhat compensate for this.

7

Robert Klemme wrote:

I guess what confused me is that I bring to Ruby the notion that if a symbol causes something to happen, as "?" clearly does in my example (it produces the integer code for a character),

In this case nothing "happens" (see below). It is just another syntactical way to express an integer constant.

its' an operator or method. I'm not unfamiliar with linguistics. I don't see the equivalence between a string delimiter, or a character that signals the beginning of a symbol, and a symbol that is actually productive of something. Makes no sense to me at all, in fact. Maybe I'm failing to grasp some transformative concept which changes how things work in the Ruby world.

Seeing operators as methods makes some real sense, on the other hand. "=" produces something out of what it's given. It doesn't just change meaning, which is all that happens with quotes or ":". That's the distinction I'm seeing, and it seems valid.

I am not sure what you mean by "change meaning". Can you elaborate?

Meaning, in linguistics as I understand it, is inherent, contextural, or both. "a" is inherently a character in the basic Roman alphabet used by many European languages. As a representation of a sound, its meaning changes depending upon context. Consider "pat" and "pate". Adding the suffix "e" to the consonant+vowel+consonant construction changes the sound of the "a". In Ruby, customer_count (a variable) and "customer_count" (a string) don't mean the same thing, because the latter has a prefix and suffix of the " symbol. Contextural meaning production is occurring. Ditto for the use of ":" in symbols. The """ and the ":" are context indicators, but produce nothing. Not so for "+" in "a+b" or, I think "?a". Something is emitted, here. That goes well beyond mere contextual meaning production, which is also occurring, to be sure. I hope I'm being clear enough here.

Btw, note that there are two types of "=":

1. variable assignment as in "a = 123",

2. method call that mimics variable assignment as in "a.foo = 456".

Both evaluate always to the right hand side - but this is about the only commonality. Semantics of type 1 is fixed while you are free to implement type 2 in any way you like, e.g.

irb(main):005:0> a = Object.new
=> #<Object:0x7ff79980>
irb(main):006:0> def a.foo=(x) puts "bah!" end
=> nil
irb(main):007:0> a.foo = 456
bah!
=> 456
irb(main):008:0>

Thanks! That's fun, and interesting.

I'll keep thinking about it.

Basically "?a" is just another textual representation for the integer 97 which happens to be the ASCII code of character "a". It differs from string delimiters because the expression "foo" is actually a String constructor - it will create a new String object whenever it is executed:

irb(main):003:0> 3.times { x = "foo"; puts x, x.object_id }

A "textural representation" that PRODUCES something? Doesn't make sense to me. Representations simply represent. They "stand for" something else, and do no more.

Now, I can define method X, and then say that X "stands for" the method specification. That's true, from the compiler's point of view, among other things. But to say no more is misleading. "X" can be called, at which point it ACTS. "Tom" - my name can also be called, but when one does so IT doesn't spring into action at all. I might, but IT cannot.

There is, I therefore assert, a fundamental difference here, and I think you're seeing that two different things have the same feature and concluding from that that they are the same sort of thing. That argument is obviously flawed. "It's incomplete, as my "Tom" demonstrations shows. You say "Basically "?a" is just another textual representation for..." - I'd agree, except for that "just". That single word makes the statement incorrect, to my mind, and also reveals the problem I'm addressing.

foo
1073491940
foo
1073492040
foo
1073491880
=> 3
irb(main):004:0>

[...]

Thanks for your thoughts. I appreciate your sharing them.

t.

···

On 04.01.2009 14:48, Tom Cloyd wrote:

--

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd, MS MA, LMHC - Private practice Psychotherapist
Bellingham, Washington, U.S.A: (360) 920-1226
<< tc@tomcloyd.com >> (email)
<< TomCloyd.com >> (website) << sleightmind.wordpress.com >> (mental health weblog)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

8

Thanks for the education! I wasn't aware of this. There was
definitively an illegal PDF version offered (even here) in the past.

Kind regards

robert

···

2009/1/4 Michael Guterl <mguterl@gmail.com>:

On Sun, Jan 4, 2009 at 9:48 AM, Robert Klemme > <shortcutter@googlemail.com> wrote:

As far as I know there is no legal electronic version of the second edition
so you better double check whether you are allowed to use this. You could
as well buy the book to somewhat compensate for this.

Pragmatic Bookshelf: By Developers, For Developers

There definitely is an option to select a PDF version for purchase.

--
remember.guy do |as, often| as.you_can - without end

9

Robert Klemme wrote:

I guess what confused me is that I bring to Ruby the notion that if a symbol causes something to happen, as "?" clearly does in my example (it produces the integer code for a character),

In this case nothing "happens" (see below). It is just another syntactical way to express an integer constant.

its' an operator or method. I'm not unfamiliar with linguistics. I don't see the equivalence between a string delimiter, or a character that signals the beginning of a symbol, and a symbol that is actually productive of something. Makes no sense to me at all, in fact. Maybe I'm failing to grasp some transformative concept which changes how things work in the Ruby world.

Seeing operators as methods makes some real sense, on the other hand. "=" produces something out of what it's given. It doesn't just change meaning, which is all that happens with quotes or ":". That's the distinction I'm seeing, and it seems valid.

I am not sure what you mean by "change meaning". Can you elaborate?

Meaning, in linguistics as I understand it,

Unfortunately we're in the domain of programming languages which is quite different from linguistics. Part of my difficulty understanding you is probably caused by the fact that you seem to try to tackle problems of computer languages with tools (terms and theories) from a complete different domain. While both domains happen to share terms I think it is not helpful to try to apply linguistics (i.e. the science of natural languages) to computer languages. It may actually hinder understanding.

Ditto for the use of ":" in symbols. The """ and the ":" are context indicators, but produce nothing. Not so for "+" in "a+b" or, I think "?a".

Actually I believe that "?a" does not belong in the same category as "a+b" because, as Dave and I have pointed out already, "?a" is just another way to write the integer 97 ("97", "0141" and "0x61" are other ways to write 97).

I'll keep thinking about it.

Basically "?a" is just another textual representation for the integer 97 which happens to be the ASCII code of character "a". It differs from string delimiters because the expression "foo" is actually a String constructor - it will create a new String object whenever it is executed:

A "textural representation" that PRODUCES something?

I did not mention "produce" in the context of - where do you take that from?

Doesn't make sense to me. Representations simply represent. They "stand for" something else, and do no more.

Yes, all these character sequences ("?a", "97", "0141" and "0x61") represent the Fixnum instance which in term is a technical representation (i.e. data) of the (mathematical) number 97. If an expression of a Ruby program is evaluate that contains any of these tokens, the token (which is an expression as well, more precisely a constant expression) evaluates to the said Fixnum instance.

Now, I can define method X, and then say that X "stands for" the method specification. That's true, from the compiler's point of view, among other things. But to say no more is misleading. "X" can be called, at which point it ACTS. "Tom" - my name can also be called, but when one does so IT doesn't spring into action at all. I might, but IT cannot.

It does, as my example demonstrated. Opposed to "?a" "'Tom'" is not a constant expression but a String constructor. (See my example below which I took the liberty to put together again.)

There is, I therefore assert, a fundamental difference here, and I think you're seeing that two different things have the same feature and

Actually I pointed out that "?a" and "'Tom'" are not the same sort of thing. (On a certain level they are of course: both are valid syntax constructs (tokens) of the Ruby language. Actually, both are even expressions which happen do have different semantics - constant vs. non constant).

concluding from that that they are the same sort of thing.

Again, I do not know where you take that from. This is certainly nothing I have intended to convey.

That argument is obviously flawed. "It's incomplete, as my "Tom" demonstrations shows. You say "Basically "?a" is just another textual representation for..." - I'd agree, except for that "just". That single word makes the statement incorrect, to my mind, and also reveals the problem I'm addressing.

Not at all. There are many ways to write down 97 in a Ruby program (as I have shown above). In that sense "?a" is just another way to write it down (assuming "97" is considered the most commonly used way).

irb(main):003:0> 3.times { x = "foo"; puts x, x.object_id }
foo
1073491940
foo
1073492040
foo
1073491880
=> 3
irb(main):004:0>

Cheers

  robert

···

On 05.01.2009 03:38, Tom Cloyd wrote:

On 04.01.2009 14:48, Tom Cloyd wrote:

--
remember.guy do |as, often| as.you_can - without end

10

Robert Klemme wrote:

  

As far as I know there is no legal electronic version of the second edition
so you better double check whether you are allowed to use this. You could
as well buy the book to somewhat compensate for this.

Pragmatic Bookshelf: By Developers, For Developers

There definitely is an option to select a PDF version for purchase.
    
Thanks for the education! I wasn't aware of this. There was
definitively an illegal PDF version offered (even here) in the past.

Kind regards

robert

I wouldn't engage with a illegitimate copy - isn't playing fair at all. I completely believe in "intellectual property". My copy certainly WAS purchased from the Pragmatic Programmers website. Marvelous book, of course, but it does not, to my best knowledge, contain a complete list of operators. I've done everything I can think of to find it. I seriously doubt that it's there. The precedence table on 324 is about precedence, and not a primary presentation of operators and their function.

More incredibly, "operator" is not even in the index as a primary index term - in the sense that I'm accustomed to the use of the word in programming languages.

Here's the entry that IS there:

Operator
   as method call 82, 335
   precedence 324

To me, this is simply strange. Dave Thomas, whose contributions to the Ruby community are inestimably valuable, may see this thread, in which case I hope he comments. I've been completely bewildered about this for a long time.

t.

···

2009/1/4 Michael Guterl <mguterl@gmail.com>:

On Sun, Jan 4, 2009 at 9:48 AM, Robert Klemme >> <shortcutter@googlemail.com> wrote:

--

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd, MS MA, LMHC - Private practice Psychotherapist
Bellingham, Washington, U.S.A: (360) 920-1226
<< tc@tomcloyd.com >> (email)
<< TomCloyd.com >> (website) << sleightmind.wordpress.com >> (mental health weblog)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

11

Not sure if the following list is comprehensive, but useful nonetheless...

...you'll find the first use of ? under Reserved words.

Todd

···

On Sun, Jan 4, 2009 at 5:06 PM, Tom Cloyd <tomcloyd@comcast.net> wrote:

Robert Klemme wrote:

2009/1/4 Michael Guterl <mguterl@gmail.com>:

On Sun, Jan 4, 2009 at 9:48 AM, Robert Klemme >>> <shortcutter@googlemail.com> wrote:

As far as I know there is no legal electronic version of the second
edition
so you better double check whether you are allowed to use this. You
could
as well buy the book to somewhat compensate for this.

Pragmatic Bookshelf: By Developers, For Developers

There definitely is an option to select a PDF version for purchase.

Thanks for the education! I wasn't aware of this. There was
definitively an illegal PDF version offered (even here) in the past.

Kind regards

robert

I wouldn't engage with a illegitimate copy - isn't playing fair at all. I
completely believe in "intellectual property". My copy certainly WAS
purchased from the Pragmatic Programmers website. Marvelous book, of course,
but it does not, to my best knowledge, contain a complete list of operators.
I've done everything I can think of to find it. I seriously doubt that it's
there. The precedence table on 324 is about precedence, and not a primary
presentation of operators and their function.

More incredibly, "operator" is not even in the index as a primary index term
- in the sense that I'm accustomed to the use of the word in programming
languages.

Here's the entry that IS there:

Operator
as method call 82, 335
precedence 324

To me, this is simply strange. Dave Thomas, whose contributions to the Ruby
community are inestimably valuable, may see this thread, in which case I
hope he comments. I've been completely bewildered about this for a long
time.

12

I wouldn't engage with a illegitimate copy - isn't playing fair at all. I completely believe in "intellectual property". My copy certainly WAS purchased from the Pragmatic Programmers website.

And we appreciate that :slight_smile:

Marvelous book, of course, but it does not, to my best knowledge, contain a complete list of operators. I've done everything I can think of to find it. I seriously doubt that it's there. The precedence table on 324 is about precedence, and not a primary presentation of operators and their function.

Which methods/operators are missing from the precedence table? I'll add them.

?<char>, tho' is not an operator, any more than the quote is in "cat" or the slash in /cat/. In all three cases they're simply syntax for literals.

Here's the entry that IS there:

Operator
as method call 82, 335
precedence 324

To me, this is simply strange. Dave Thomas, whose contributions to the Ruby community are inestimably valuable, may see this thread, in which case I hope he comments. I've been completely bewildered about this for a long time.

Let me know what to add, and I'll definitely consider adding it. But be aware that I don't really consider the concept of "an operator" to be particular primary in the description of the language, because in Ruby there'll always be debate about just what _is_ an operator. The precedence table is my best take on it—I derived it from the parser, and I believe it contains every operator-like thing I could find in there. But, as I said, I'd love to hear suggestions, as the third edition is nearing completion, and I'm always open for ideas to make it better.

Cheers

Dave

···

On Jan 4, 2009, at 5:06 PM, Tom Cloyd wrote:

13

I meant beneath Reserved words. It's actually classified under Types/Numbers.

Todd

···

On Sun, Jan 4, 2009 at 5:36 PM, Todd Benson <caduceass@gmail.com> wrote:

On Sun, Jan 4, 2009 at 5:06 PM, Tom Cloyd <tomcloyd@comcast.net> wrote:

Robert Klemme wrote:

2009/1/4 Michael Guterl <mguterl@gmail.com>:

On Sun, Jan 4, 2009 at 9:48 AM, Robert Klemme >>>> <shortcutter@googlemail.com> wrote:

As far as I know there is no legal electronic version of the second
edition
so you better double check whether you are allowed to use this. You
could
as well buy the book to somewhat compensate for this.

http://www.pragprog.com/titles/ruby/programming-ruby

There definitely is an option to select a PDF version for purchase.

Thanks for the education! I wasn't aware of this. There was
definitively an illegal PDF version offered (even here) in the past.

Kind regards

robert

I wouldn't engage with a illegitimate copy - isn't playing fair at all. I
completely believe in "intellectual property". My copy certainly WAS
purchased from the Pragmatic Programmers website. Marvelous book, of course,
but it does not, to my best knowledge, contain a complete list of operators.
I've done everything I can think of to find it. I seriously doubt that it's
there. The precedence table on 324 is about precedence, and not a primary
presentation of operators and their function.

More incredibly, "operator" is not even in the index as a primary index term
- in the sense that I'm accustomed to the use of the word in programming
languages.

Here's the entry that IS there:

Operator
as method call 82, 335
precedence 324

To me, this is simply strange. Dave Thomas, whose contributions to the Ruby
community are inestimably valuable, may see this thread, in which case I
hope he comments. I've been completely bewildered about this for a long
time.

Not sure if the following list is comprehensive, but useful nonetheless...

Ruby | zenspider.com | by ryan davis

...you'll find the first use of ? under Reserved words.

14

Dave Thomas wrote:

I wouldn't engage with a illegitimate copy - isn't playing fair at all. I completely believe in "intellectual property". My copy certainly WAS purchased from the Pragmatic Programmers website.

And we appreciate that :slight_smile:

Marvelous book, of course, but it does not, to my best knowledge, contain a complete list of operators. I've done everything I can think of to find it. I seriously doubt that it's there. The precedence table on 324 is about precedence, and not a primary presentation of operators and their function.

Which methods/operators are missing from the precedence table? I'll add them.

Oh, man...now I've going to have to get rigorous in my consideration of the topic, aren't I? Heh heh. Probably about time, too. So...I'll go on a search and recover mission, if only for my own edification. If I find anything, I'll report back. Probably will anyway.

But to me the basic problem with the p. 324 table, as I suggested, is that its subject is about precedence - a completely legitimate subject - and is NOT primarily a presentation of operators with an explication of what they DO. THAT's what I looked for practically the first time I encountered the book, and have been looking for ever since.

As I just said in a response to Todd's post, The best presentation of Ruby operators I've yet found (looking mostly on the web) is at http://www.tutorialspoint.com/ruby/ruby_operators.htm\. THIS is the sort of presentation I expect in ANY language book. It seems fully as basic as a definition of what a variable is and how one works, along with all the other basic building blocks of the language. Languages are build up from primitives - analogous to the premises in a logical argument. If the premises are unclear, the argument cannot go forward, and that's been exactly my experience with Ruby. I encounter some strange operator - like the "?" encounter I had last night, and I want to go read about this thing. After well over 30 minutes search through quite a range of material, I had found next to nothing. That's nuts, to me.

?<char>, tho' is not an operator, any more than the quote is in "cat" or the slash in /cat/. In all three cases they're simply syntax for literals.

I'm still struggling here. When syntax is productive of something, I see an operation occurring - active functionality. I'm perfectly content to see operators as methods. Syntactically there's no problem that I can detect in that construct. But to call "?<char>" a literal...well, I must go study up a bit on this. To me, "abcdef" is a literal. '\n' is too. "\n" is a literal if you allow meaning to be contexturalized, which should be no problem. As input to various processes, '\n' and "\n" have different outcomes, but by themselves produce nothing. "?<char>" produces something. Irb makes this clear (to me):

irb(main):006:0> puts '\n'
\n
=> nil
irb(main):007:0> puts "\n"

=> nil
irb(main):008:0> puts ?\n
10
=> nil
irb(main):009:0>

If in "-1", we consider the "-" to be an operator, then I see an exact parallel to the function of "?a".

It's not the same as what quote marks do - THEY provide context, and change meaning. Operators PRODUCE something. They "operate". It seems very simple. If I'm making a fundamental error, I'd be very grateful to have it point out to me. I highly value arriving at a place of clear understanding, when it's possible.

Here's the entry that IS there:

Operator
as method call 82, 335
precedence 324

To me, this is simply strange. Dave Thomas, whose contributions to the Ruby community are inestimably valuable, may see this thread, in which case I hope he comments. I've been completely bewildered about this for a long time.

Let me know what to add, and I'll definitely consider adding it. But be aware that I don't really consider the concept of "an operator" to be particular primary in the description of the language, because in Ruby there'll always be debate about just what _is_ an operator. The precedence table is my best take on it—I derived it from the parser, and I believe it contains every operator-like thing I could find in there. But, as I said, I'd love to hear suggestions, as the third edition is nearing completion, and I'm always open for ideas to make it better.

Dave, the function descriptions in this table are too minimal, I think. Not enough for a beginner - not for this one, anyway, and Ruby is at least the fifth language I've dug around in. In all the others (Fortran, APL, Pascal/Delphi, Basic, Python, a tabular presentation of the operators the languages recognizes was shown to me EARLY, and I always found it highly informative, since it told me what I could DO with primitive units of data. Seems like an excellent starting point for more complex topics.

Your comment about the operator concept is very interesting. I think that it would be really, truly helpful for you to lay out Ruby more from the ground up - shown me clearly it fundamental components, beginning with the 'primitives' - reserved words are primitives, and so are variables, etc. Operators, too, I'd say. From these things we build other things, like data containers - meta-primitives, if you will.

I have always been greatly bewildered that in your book (I'm looking at 2nd edition PDF again), we're into arrays and hashes on p. 14 and I have to encounter anything substantive about operators - those things that glue together variables and constants so that we can form expressions.

Maybe I just bring the wrong mindset to Ruby. I'm willing to be told that, if I'm shown a better one. However, I'd bet that a great many of us did study grammar in school (remember "parts of speech"?), and some of us have studied linguistics, and many of us have at least dabbled in other languages. There are conventions about how we talk about human communication, in general, computer languages included. When I come to a presentation of a new language, the first thing I look for is the familiar. I want some secure ground I can readily claim ("oh, I know what THAT is!"), so that I can move from that out into the wilderness of new territory.

For me that secure ground in anything of linguistic nature will always be the "meme" concept - in formal linguistics this would be an analysis of phonemes (aural primitives - irreducible units of sound) and graphemes (irreducible units of sound representation), etc. None of these concepts are perfect, but they ARE plainly very useful. The basic model is very familiar: premises (or data, if you will) + logic (legal operations) = argument (or assertion or expression). This model is a part of our cultural memory, as it were.

I want someone to show me the pieces of something, and how they go together. I don't want complexity until I grasp the little things, at least at the conceptual, if not the operational/functional, level. I love the highly informative narratives you lay out in your book, about all manner of topics. But I'm bewildered that in Ch. 2, the first real content about the language per se, we're right away talking about OO programming. I don't yet have the pieces of the language, at this point. How can I think about how OO programming is done in Ruby. It's as if I'm being asked to walk on air. Obviously, this is just one man's reaction, and I full well know you'll see it that way. No problem.

If ANY of this is of any use to you in any way, it's worth the time it took for me to lay it out. I don't expect a response. I'm only hoping that maybe some of my perspective might be useful to you. If you get a chance to read it at all, I'm content simply with that.

3rd edition coming out soon, eh? I have to say that one of the great joys of being in the Ruby community is that the future looks so interesting - and this is one piece of it we can all look forward to. I'm sure you'll have my money soon enough, and it'll be a bargain! Best of luck with your work on this project, and my thanks as well.

Thanks...

Tom

···

On Jan 4, 2009, at 5:06 PM, Tom Cloyd wrote:

Cheers

Dave

--

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd, MS MA, LMHC - Private practice Psychotherapist
Bellingham, Washington, U.S.A: (360) 920-1226
<< tc@tomcloyd.com >> (email)
<< TomCloyd.com >> (website) << sleightmind.wordpress.com >> (mental health weblog)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

15

Todd Benson wrote:

  

Robert Klemme wrote:
      

As far as I know there is no legal electronic version of the second
edition
so you better double check whether you are allowed to use this. You
could
as well buy the book to somewhat compensate for this.

Pragmatic Bookshelf: By Developers, For Developers

There definitely is an option to select a PDF version for purchase.

Thanks for the education! I wasn't aware of this. There was
definitively an illegal PDF version offered (even here) in the past.

Kind regards

robert

I wouldn't engage with a illegitimate copy - isn't playing fair at all. I
completely believe in "intellectual property". My copy certainly WAS
purchased from the Pragmatic Programmers website. Marvelous book, of course,
but it does not, to my best knowledge, contain a complete list of operators.
I've done everything I can think of to find it. I seriously doubt that it's
there. The precedence table on 324 is about precedence, and not a primary
presentation of operators and their function.

More incredibly, "operator" is not even in the index as a primary index term
- in the sense that I'm accustomed to the use of the word in programming
languages.

Here's the entry that IS there:

Operator
as method call 82, 335
precedence 324

To me, this is simply strange. Dave Thomas, whose contributions to the Ruby
community are inestimably valuable, may see this thread, in which case I
hope he comments. I've been completely bewildered about this for a long
time.
      

Not sure if the following list is comprehensive, but useful nonetheless...

Ruby | zenspider.com | by ryan davis

...you'll find the first use of ? under Reserved words.
    
I meant beneath Reserved words. It's actually classified under Types/Numbers.

Todd

Definitely a useful page, about which I've known for some time. Outstanding for its brevity. Limited at times, because if you don't already know what he's talking about, you're not much enlightened by the page content at all. Two examples:
* "Here docs" (right above "Symbols"). I know what this is, and I STILL cannot quickly make sense of the notes on the pages about this
functionality.
* Operators - has the same problem as Thomas' book: precedence is shown, but basic operator function appears to be assumed to be known. I'm sorry, but I wasn't born with that knowledge. Look at this line: <=> == === != =~ !~

There's some very strange stuff there. I'm only certain about the function of ONE of these, and I've been using Ruby for probably over a year an a half (remember - I'm an intermittent weekend programmer, of necessity).

All that said, I find Ryan Davis's work simply amazing, and I'm very grateful for his presence in the world. His "Ruby QuickRef" at Ruby | zenspider.com | by ryan davis is likely supposed to be more cheat sheet than brief tutorial, to be fair.

The best presentation of Ruby operators I've yet found (looking mostly on the web) is at http://www.tutorialspoint.com/ruby/ruby_operators.htm\. Also good are those at Ruby Operators - Techotopia and Dev Trademarks - Furm. So, this stuff can be found, but you have to go looking for it.

t.

···

On Sun, Jan 4, 2009 at 5:36 PM, Todd Benson <caduceass@gmail.com> wrote:

On Sun, Jan 4, 2009 at 5:06 PM, Tom Cloyd <tomcloyd@comcast.net> wrote:

2009/1/4 Michael Guterl <mguterl@gmail.com>:

On Sun, Jan 4, 2009 at 9:48 AM, Robert Klemme >>>>> <shortcutter@googlemail.com> wrote:

--

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd, MS MA, LMHC - Private practice Psychotherapist
Bellingham, Washington, U.S.A: (360) 920-1226
<< tc@tomcloyd.com >> (email)
<< TomCloyd.com >> (website) << sleightmind.wordpress.com >> (mental health weblog)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

16

I think you're over thinking it. For example, in

   0x10

which is 16 in decimal, is the '0x' part an operator? If you say no (which I hope you do) then why would you say that ? is an operator in ?a (which is 97). They are simply characters that are used by the lexical analysis phase to help interpret what follows. No executable code is generated: there's no operation.

Similarly, in your example of -1, the - sign is NOT an operator. It simply means that when parsing the constant, a negative value is required. Again, so executable code is generated. That's the difference between

    a = -1 # -1 is a constant, no operator
    b = -a # the unary minus operator (-@) is invoked on the object referenced by a

Cheers

Dave

···

On Jan 4, 2009, at 8:14 PM, Tom Cloyd wrote:

I'm still struggling here. When syntax is productive of something, I see an operation occurring - active functionality. I'm perfectly content to see operators as methods. Syntactically there's no problem that I can detect in that construct. But to call "?<char>" a literal...well, I must go study up a bit on this. To me, "abcdef" is a literal. '\n' is too. "\n" is a literal if you allow meaning to be contexturalized, which should be no problem. As input to various processes, '\n' and "\n" have different outcomes, but by themselves produce nothing. "?<char>" produces something. Irb makes this clear (to me):

irb(main):006:0> puts '\n'
\n
=> nil
irb(main):007:0> puts "\n"

=> nil
irb(main):008:0> puts ?\n
10
=> nil
irb(main):009:0>

If in "-1", we consider the "-" to be an operator, then I see an exact parallel to the function of "?a".

It's not the same as what quote marks do - THEY provide context, and change meaning. Operators PRODUCE something. They "operate". It seems very simple. If I'm making a fundamental error, I'd be very grateful to have it point out to me. I highly value arriving at a place of clear understanding, when it's possible

17

Dave Thomas wrote:

I'm still struggling here. When syntax is productive of something, I see an operation occurring - active functionality. I'm perfectly content to see operators as methods. Syntactically there's no problem that I can detect in that construct. But to call "?<char>" a literal...well, I must go study up a bit on this. To me, "abcdef" is a literal. '\n' is too. "\n" is a literal if you allow meaning to be contexturalized, which should be no problem. As input to various processes, '\n' and "\n" have different outcomes, but by themselves produce nothing. "?<char>" produces something. Irb makes this clear (to me):

irb(main):006:0> puts '\n'
\n
=> nil
irb(main):007:0> puts "\n"

=> nil
irb(main):008:0> puts ?\n
10
=> nil
irb(main):009:0>

If in "-1", we consider the "-" to be an operator, then I see an exact parallel to the function of "?a".

It's not the same as what quote marks do - THEY provide context, and change meaning. Operators PRODUCE something. They "operate". It seems very simple. If I'm making a fundamental error, I'd be very grateful to have it point out to me. I highly value arriving at a place of clear understanding, when it's possible

I think you're over thinking it. For example, in

  0x10

which is 16 in decimal, is the '0x' part an operator? If you say no (which I hope you do) then why would you say that ? is an operator in ?a (which is 97). They are simply characters that are used by the lexical analysis phase to help interpret what follows. No executable code is generated: there's no operation.

Similarly, in your example of -1, the - sign is NOT an operator. It simply means that when parsing the constant, a negative value is required. Again, so executable code is generated. That's the difference between

   a = -1 # -1 is a constant, no operator
   b = -a # the unary minus operator (-@) is invoked on the object referenced by a

Cheers

Dave

Thanks, Dave. This helps. You have a genuine knack for simple clear statement and example. Thanks for the dialog!

t.

···

On Jan 4, 2009, at 8:14 PM, Tom Cloyd wrote:

--

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd, MS MA, LMHC - Private practice Psychotherapist
Bellingham, Washington, U.S.A: (360) 920-1226
<< tc@tomcloyd.com >> (email)
<< TomCloyd.com >> (website) << sleightmind.wordpress.com >> (mental health weblog)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

18

Tom Cloyd wrote:

* Operators - has the same problem as Thomas' book: precedence is shown,
but basic operator function appears to be assumed to be known. I'm
sorry, but I wasn't born with that knowledge. Look at this line: <=> ==
=== != =~ !~

There aren't really any operators in Ruby. Ruby simply maps all these to
method calls (*), so to find out what the operator *means* you have to
look at the class documentation for that method.

e.g.
  str1 = "foo"
  str2 = "bar"
  p str1 <=> str2

Here you need to look at the documentation for String#<=> to see what it
means. Yes, "<=>" is a legitimate method name :slight_smile: You can also call it
like this:

  str1.<=>(str2)

The only real purpose of the operator precedence table is to show how
expressions containing multiple operators are resolved. e.g.

   1 + 2 * 3

# same as:
# 1 + (2 * 3)

# same as:
# 1.+(2.*(3))

Note that even the methods '+' and '*' do not necessarily have anything
to do with addition or multiplication. It's entirely dependent on the
target object how it implements these methods. For example:

   "123" * 3 # => "123123123"

HTH,

Brian.

(*) The exceptions in ruby 1.8 are != and !~, which invoke the methods
== and =~ but then negate the result

···

--
Posted via http://www.ruby-forum.com/\.

19

Tom Cloyd wrote:

<=> == === != =~ !~

There's some very strange stuff there. I'm only certain about the
function of ONE of these

As has already been mentionen, they are methods and do whatever they're
defined to do. That being said here's what their intended function is:
<=> should be defined by any class that includes the Comaparable module. It
should return -1,0 or 1 depending on whether the receiver (the lhs) is less
than, equal to or greater than the argument (the rhs).

== should return true if (and only if) the lhs is equal to the rhs (where the
definition of equality depends on the class). Object#== is defined to return
true if two objects are actually the same object (i.e. both the receiver and
the argument are actually references to the same location in memory).

=== defines when an object "matches" another object in a case-when-statement
(or the grep method). For example in the following code, the puts will be
executed if y === x is true:
case x
when y
puts "tada"
end
Object#=== is defined to return the same as ==, Regexp#=== is defined to
return true if the regexp matches a string and Range#=== is defined to
return true if a given object is inside the range.

!= is simply the negation of ==, so x != y is equivalent to !(x == y)

Regexp#=~, which takes a string as an argument, and String#=~, which takes a
regex as an argument, return the index where the regex matches the string or
nil if the regex does not match the string.

All of those are documented in the API docs btw.

HTH,
Sebastian

···

--
Jabber: sepp2k@jabber.org
ICQ: 205544826

20

Brian Candler wrote:

Tom Cloyd wrote:
  

* Operators - has the same problem as Thomas' book: precedence is shown,
but basic operator function appears to be assumed to be known. I'm
sorry, but I wasn't born with that knowledge. Look at this line: <=> ==
=== != =~ !~
    
There aren't really any operators in Ruby. Ruby simply maps all these to method calls (*), so to find out what the operator *means* you have to look at the class documentation for that method.

e.g.
  str1 = "foo"
  str2 = "bar"
  p str1 <=> str2

Here you need to look at the documentation for String#<=> to see what it means. Yes, "<=>" is a legitimate method name :slight_smile: You can also call it like this:

  str1.<=>(str2)

The only real purpose of the operator precedence table is to show how expressions containing multiple operators are resolved. e.g.

   1 + 2 * 3

# same as:
# 1 + (2 * 3)

# same as:
# 1.+(2.*(3))

Note that even the methods '+' and '*' do not necessarily have anything to do with addition or multiplication. It's entirely dependent on the target object how it implements these methods. For example:

   "123" * 3 # => "123123123"

HTH,

Brian.

(*) The exceptions in ruby 1.8 are != and !~, which invoke the methods == and =~ but then negate the result
  

Ah...This is actually very helpful. I think we have a language problem, though. If there are no operators in Ruby, we should call all those methods with non-alphabetical names something besides "operators". I know...we get USED to nonsense. Every try to learn to spell English? I still don't do it well. But we live with it. Here, I think we need to change not only the computer language (experts appear to be evolving it quite nicely, and I'm happy just to watch), but the human language we use to talk about it. I guess this particularly matters to me because I write a lot - daily. I virtually live at a keyboard. Verbal coherence matters a lot to me, even if I can't always achieve it.

OK, let's assume we have the language problem solved (ha! - not likely to happen). Nann-methods == "non-alphabetically-named methods". The core problem still remains - what the blazes does that THING do/mean/produce...??? Here's were your post is helpful to me. I will follow your suggestion. I feel better already, because for the first time in a long time [a] I understand my confusion, AND [2] what to do about it. A whole new world.

Dave...if you're still listening...it'd be lovely if some clarification along the lines of this post made it into your 3rd edition - that's what I find myself thinking, anyway.

Thanks to all...for hanging with me on this one.

Thanks.

T.

···

--

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd, MS MA, LMHC - Private practice Psychotherapist
Bellingham, Washington, U.S.A: (360) 920-1226
<< tc@tomcloyd.com >> (email)
<< TomCloyd.com >> (website) << sleightmind.wordpress.com >> (mental health weblog)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~