Calling a function with undefined symbol

Calling a function with undefined symbol

[Heime]

Have written a function named "mbcomplt" that takes a symbol.

(defun mbcomplt (sgnl)
  "DOC."

(if (eq sgnl 'go)
    (message "go")
  (message "nogo")))

I can call this function with

(mbcomplt 'go)

This even though the symbol "'go" does not exist.  What is happening here?

My understanding about symbols is that I can make a symbol using

(defvar seqr 3)  ; set symbol seqr with value 3
(setq seqr 5)    ; set symbol seqr with value 5

(symbol-value 'seqr) ; get value of symbol (returns 5)
(symbol-name 'seqr)  ; get name of symbol (returns "seqr")

Re: Calling a function with undefined symbol

[Jean Abou Samra]

Le 31/10/2022 à 09:15, Heime a écrit :
> Have written a function named "mbcomplt" that takes a symbol.
>
> (defun mbcomplt (sgnl)
>    "DOC."
>
> (if (eq sgnl 'go)
>      (message "go")
>    (message "nogo")))
>
> I can call this function with
>
> (mbcomplt 'go)
>
> This even though the symbol "'go" does not exist.  What is happening here?
>
> My understanding about symbols is that I can make a symbol using
>
> (defvar seqr 3)  ; set symbol seqr with value 3
> (setq seqr 5)    ; set symbol seqr with value 5
>
> (symbol-value 'seqr) ; get value of symbol (returns 5)
> (symbol-name 'seqr)  ; get name of symbol (returns "seqr")




It may be helpful to view symbols as some kind of modified
strings. The symbol 'go is an amorphous object, just a name
for something. You can use it as a name, or as a discrete
key, or whatever. You can also "define" it, or rather
define a variable with its name, and then you are able
to use symbol-value on it, but the symbol object itself is
a *potential* name and does not need to have a variable with
its name defined in order to be used.

Jean

Re: Calling a function with undefined symbol

[tomas]

[-- Attachment #1: Type: text/plain, Size: 1591 bytes --]

On Mon, Oct 31, 2022 at 08:15:32AM +0000, Heime wrote:
> 
> Have written a function named "mbcomplt" that takes a symbol.

The function takes an argument, that's what it does.

> (defun mbcomplt (sgnl)
>   "DOC."

sgnl here is a variable. Or a variable name. Depending on how
you look at it.

> (if (eq sgnl 'go)
>     (message "go")
>   (message "nogo")))

Now you are comparing sgnl's /value/ to something (which happens
to be a symbol).

> I can call this function with
> 
> (mbcomplt 'go)

Yes. You could call it also with (mbcomplt 23). Or (mbcomplt "This is not a pipe").
So?

> This even though the symbol "'go" does not exist.  What is happening here?

As soon as you do 'go, the symbol exists.

> My understanding about symbols is that I can make a symbol using
> 
> (defvar seqr 3)  ; set symbol seqr with value 3

No, no. You are making a variable. Which is a binding of a symbol
to a value.

> (setq seqr 5)    ; set symbol seqr with value 5

No, you set the variable's value to 5. The symbol is not "set" to
anything. It is an immutable thing, as far as Lisp is concerned.

The symbol 'seqr is just taking the job of variable name here.
Like you are called Heime (or whatever it is this week). That
name is not you.

> (symbol-value 'seqr) ; get value of symbol (returns 5)
> (symbol-name 'seqr)  ; get name of symbol (returns "seqr")

Symbol-value returns the value associated to the symbol. The
association itself is called a "variable".

Please go re-read the manuals. You have got all of it confused :)

Cheers
-- 
t

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

Re: Calling a function with undefined symbol

[Heime]

------- Original Message -------
On Monday, October 31st, 2022 at 8:42 AM, <tomas@tuxteam.de> wrote:


> On Mon, Oct 31, 2022 at 08:15:32AM +0000, Heime wrote:
> 
> > Have written a function named "mbcomplt" that takes a symbol.
> 
> 
> The function takes an argument, that's what it does.
> 
> > (defun mbcomplt (sgnl)
> > "DOC."
> 
> 
> sgnl here is a variable. Or a variable name. Depending on how
> you look at it.
> 
> > (if (eq sgnl 'go)
> > (message "go")
> > (message "nogo")))
> 
> 
> Now you are comparing sgnl's /value/ to something (which happens
> to be a symbol).
> 
> > I can call this function with
> > 
> > (mbcomplt 'go)
> 
> 
> Yes. You could call it also with (mbcomplt 23). Or (mbcomplt "This is not a pipe").
> So?
> 
> > This even though the symbol "'go" does not exist. What is happening here?
> 
> 
> As soon as you do 'go, the symbol exists.
> 
> > My understanding about symbols is that I can make a symbol using
> > 
> > (defvar seqr 3) ; set symbol seqr with value 3
> 
> 
> No, no. You are making a variable. Which is a binding of a symbol
> to a value.

Ok, I make a variable.  What happens after that?  A symbol 'seqr is made too?
 
> > (setq seqr 5) ; set symbol seqr with value 5
> 
> 
> No, you set the variable's value to 5. The symbol is not "set" to
> anything. It is an immutable thing, as far as Lisp is concerned.

Then, how is it that (symbol-value 'seqr) returns 5?  There seems to be 
a symbol and symbol-value returns 5.
 
> The symbol 'seqr is just taking the job of variable name here.
> Like you are called Heime (or whatever it is this week). That
> name is not you.
> 
> > (symbol-value 'seqr) ; get value of symbol (returns 5)
> > (symbol-name 'seqr) ; get name of symbol (returns "seqr")
> 
> 
> Symbol-value returns the value associated to the symbol. The
> association itself is called a "variable".
> 
> Please go re-read the manuals. You have got all of it confused :)

Yes, I have got really confused about what symbols are exactly.  Even
after reading several times.  I have never seen anybody make a symbol,
only variables.  So, what is a symbol?

Re: Calling a function with undefined symbol

[tomas]

[-- Attachment #1: Type: text/plain, Size: 1023 bytes --]

On Mon, Oct 31, 2022 at 08:57:09AM +0000, Heime wrote:
> ------- Original Message -------
> On Monday, October 31st, 2022 at 8:42 AM, <tomas@tuxteam.de> wrote:

[...]

> Then, how is it that (symbol-value 'seqr) returns 5?  There seems to be 
> a symbol and symbol-value returns 5.

Please: read the docs yourself. Others can't do that for you.

Specifically "12.7 Accessing Variable Values" in the Emacs Lisp
manual.

That a "symbol" has a "value" is just a shorthand notation for
"a symbol" has "a value" associated to it via a variable. It can
also have other things associated to it (a function, for example,
also a property list). But it needn't, it can just exist on its
own. 

> Yes, I have got really confused about what symbols are exactly.  Even
> after reading several times.  I have never seen anybody make a symbol,
> only variables.  So, what is a symbol?

'foo (spedifically this coaxes the reader into creating one. The
function behind that is intern. See there.

Cheers
-- 
t

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

Re: Calling a function with undefined symbol

[Jean Abou Samra]

> Le 31 oct. 2022 à 09:59, Heime <heimeborgia@protonmail.com> a écrit :
> 
> Yes, I have got really confused about what symbols are exactly.  Even
> after reading several times.  I have never seen anybody make a symbol,
> only variables.  So, what is a symbol?



Imagine parents who are planning to have a child and decide they will call it, say Deborah. They will say things like “I like the name Deborah better than Mary”. Those are sentences about the names themselves, not some children. The names don’t need children in order to exist as words. Then, the child is born and sentences like “Deborah weighs 3kg” (referring to the child named Deborah) start to make sense.

When you do 'symbol in Lisp, this gives you a bare name, or a “symbol”. With symbol-value, you get what value is associated with this symbol, if any (the child with that name). But you don’t need a value bound to the symbol (a child called Deborah) before you start using the symbol itself (the name “Deborah”).

Re: Calling a function with undefined symbol

[Heime]

------- Original Message -------
On Monday, October 31st, 2022 at 9:24 AM, Jean Abou Samra <jean@abou-samra.fr> wrote:


> > Le 31 oct. 2022 à 09:59, Heime heimeborgia@protonmail.com a écrit :
> > 
> > Yes, I have got really confused about what symbols are exactly. Even
> > after reading several times. I have never seen anybody make a symbol,
> > only variables. So, what is a symbol?
> 
> 
> 
> 
> Imagine parents who are planning to have a child and decide they will call it, say Deborah. They will say things like “I like the name Deborah better than Mary”. Those are sentences about the names themselves, not some children. The names don’t need children in order to exist as words. Then, the child is born and sentences like “Deborah weighs 3kg” (referring to the child named Deborah) start to make sense.
> 
> When you do 'symbol in Lisp, this gives you a bare name, or a “symbol”. With symbol-value, you get what value is associated with this symbol, if any (the child with that name). But you don’t need a value bound to the symbol (a child called Deborah) before you start using the symbol itself (the name “Deborah”).

The Lisp manual says that a symbol is an object with a name.  Then a variable (setq thevar 4) has an associated symbol 'thevar.
And the function "(defun thefun ()" also has an associated symbol "'thefun".  Furthermore one can make just a symbol, let us say 'go,
which can be passed as an argument to a function "(defun mbcomplt (arg)" using (mbcomplt 'go).  Then inside the function on can test
whether the symbol exists with (eq arg 'go).  Even though we only have a name without a values.  Is this a good basic understanding 
of a symbol?



 

Re: Calling a function with undefined symbol

[Jean Abou Samra]

> Le 31 oct. 2022 à 10:45, Heime <heimeborgia@protonmail.com> a écrit :
> 
> 
> ------- Original Message -------
> On Monday, October 31st, 2022 at 9:24 AM, Jean Abou Samra <jean@abou-samra.fr> wrote:
> 
> 
>>>> Le 31 oct. 2022 à 09:59, Heime heimeborgia@protonmail.com a écrit :
>>> 
>>> Yes, I have got really confused about what symbols are exactly. Even
>>> after reading several times. I have never seen anybody make a symbol,
>>> only variables. So, what is a symbol?
>> 
>> 
>> 
>> 
>> Imagine parents who are planning to have a child and decide they will call it, say Deborah. They will say things like “I like the name Deborah better than Mary”. Those are sentences about the names themselves, not some children. The names don’t need children in order to exist as words. Then, the child is born and sentences like “Deborah weighs 3kg” (referring to the child named Deborah) start to make sense.
>> 
>> When you do 'symbol in Lisp, this gives you a bare name, or a “symbol”. With symbol-value, you get what value is associated with this symbol, if any (the child with that name). But you don’t need a value bound to the symbol (a child called Deborah) before you start using the symbol itself (the name “Deborah”).
> 
> The Lisp manual says that a symbol is an object with a name.


This can be interpreted as the correct definition, but also as a wrong definition where a symbol is necessarily associated with a defined variable. Better said: a symbol is an object that represents a name.


> Then a variable (setq thevar 4) has an associated symbol 'thevar.
> And the function "(defun thefun ()" also has an associated symbol "'thefun".  Furthermore one can make just a symbol, let us say 'go,
> which can be passed as an argument to a function "(defun mbcomplt (arg)" using (mbcomplt 'go).


Yes. The symbol is an object like any other (5, "foo", whatever), so what you can do with an object, you can do with a symbol, including passing it as an argument to a function.


> Then inside the function on can test
> whether the symbol exists with (eq arg 'go).


That is not testing whether the symbol “exists”. It is testing whether arg is the symbol 'go.


>  Even though we only have a name without a values.  Is this a good basic understanding 
> of a symbol?


Apart from the last part, yes.

Re: Calling a function with undefined symbol

[Heime]

------- Original Message -------
On Monday, October 31st, 2022 at 9:58 AM, Jean Abou Samra <jean@abou-samra.fr> wrote:


> > Le 31 oct. 2022 à 10:45, Heime heimeborgia@protonmail.com a écrit :
> > 
> > ------- Original Message -------
> > On Monday, October 31st, 2022 at 9:24 AM, Jean Abou Samra jean@abou-samra.fr wrote:
> > 
> > > > > Le 31 oct. 2022 à 09:59, Heime heimeborgia@protonmail.com a écrit :
> > > > 
> > > > Yes, I have got really confused about what symbols are exactly. Even
> > > > after reading several times. I have never seen anybody make a symbol,
> > > > only variables. So, what is a symbol?
> > > 
> > > Imagine parents who are planning to have a child and decide they will call it, say Deborah. They will say things like “I like the name Deborah better than Mary”. Those are sentences about the names themselves, not some children. The names don’t need children in order to exist as words. Then, the child is born and sentences like “Deborah weighs 3kg” (referring to the child named Deborah) start to make sense.
> > > 
> > > When you do 'symbol in Lisp, this gives you a bare name, or a “symbol”. With symbol-value, you get what value is associated with this symbol, if any (the child with that name). But you don’t need a value bound to the symbol (a child called Deborah) before you start using the symbol itself (the name “Deborah”).
> > 
> > The Lisp manual says that a symbol is an object with a name.
> 
> 
> 
> This can be interpreted as the correct definition, but also as a wrong definition where a symbol is necessarily associated with a defined variable. Better said: a symbol is an object that represents a name.
> 
> > Then a variable (setq thevar 4) has an associated symbol 'thevar.
> > And the function "(defun thefun ()" also has an associated symbol "'thefun". Furthermore one can make just a symbol, let us say 'go,
> > which can be passed as an argument to a function "(defun mbcomplt (arg)" using (mbcomplt 'go).
> 
> 
> 
> Yes. The symbol is an object like any other (5, "foo", whatever), so what you can do with an object, you can do with a symbol, including passing it as an argument to a function.
> 
> > Then inside the function on can test
> > whether the symbol exists with (eq arg 'go).
> 
> 
> 
> That is not testing whether the symbol “exists”. It is testing whether arg is the symbol 'go.

Thank you for the clarification.  I now understand it.
 
> > Even though we only have a name without a values. Is this a good basic understanding
> > of a symbol?
> 
> 
> 
> Apart from the last part, yes.
 
Great, thank you.

Re: Calling a function with undefined symbol

[Emanuel Berg]

Heime wrote:

> I can call this function with
>
> (mbcomplt 'go)
>
> This even though the symbol "'go" does not exist.

C'mon now, it's such a common and short name, no doubt someone
in the Lisp world has defined it already!

Actually there are a lot more exotic ones that are also
symbols, e.g.

  (symbolp 'symbols-are-us) ; t

Go on, try a bunch with `symbolp' yourself, you'll be
surprised how many symbols there are out there ...

-- 
underground experts united
https://dataswamp.org/~incal

Re: Calling a function with undefined symbol

[Stefan Monnier via Users list for the GNU Emacs text editor]

>> This even though the symbol "'go" does not exist.  What is happening here?

Nitpick:

    'go

is an *expression*, not a symbol.
That expression returns a symbol, the symbol:

    go

> The symbol 'seqr is just taking the job of variable name here.

"The symbol 'foo" is like saying "The number '34".
Better write things like:

    I like the symbol `seqr` and the value `34`, as well as the
    expression `'34`.


-- Stefan

Re: Calling a function with undefined symbol

[Emanuel Berg]

Stefan Monnier via Users list for the GNU Emacs text editor wrote:

>> This even though the symbol "'go" does not exist. What is
>> happening here?
>
> Nitpick:
>
>     'go
>
> is an *expression*, not a symbol. That expression returns
> a symbol, the symbol:
>
>     go

Can't we have a list with short definitions of everything and
also a diagram with arrows like:

  An expression -> can evaluate into -> a symbol -> which can
  be bound to a value -> and that binding is then known to be
  a variable -> ...

If one is able to go thru the whole labyrinth and understand
everything then at least conceptually one has understood all
of Emacs and Lisp (Elisp).

As for practice ... this will benefit from it (as will
understanding, from practice).

>> The symbol 'seqr is just taking the job of variable
>> name here.
>
> "The symbol 'foo" is like saying "The number '34".
> Better write things like:
>
>     I like the symbol `seqr` and the value `34`, as well as
>     the expression `'34`.

Just sayin'

-- 
underground experts united
https://dataswamp.org/~incal

RE: [External] : Re: Calling a function with undefined symbol

[Drew Adams]

> > The Lisp manual says that a symbol is an object with a name.
>  
> This can be interpreted as the correct definition, but also as a wrong
> definition where a symbol is necessarily associated with a defined
> variable. Better said: a symbol is an object that represents a name.

Not really.  A symbol can "represent" any number of
things, (including a name, or multiple names - of
anything) - whatever you like.  But a symbol normally
_has_ a name, as one of its properties.


OP:

Function `make-symbol' requires a name, to create a
symbol.  Function `symbol-name' returns the name of
a symbol.

A symbol is indeed a Lisp object - an atom that has
a name and possibly other properties (value as a
variable, function definition,...).

The very first sentence of the Elisp chapter about
Symbols says this:

   A “symbol” is an object with a unique name.

And node Symbol Components tells you this:

   The print name cell _always_ holds a string, and
   cannot be changed.

   The print name cell holds the string that is the
   name of a symbol.  Since symbols are represented
   textually by their names, it is important not to
   have two symbols with the same name.[*]

Node Creating and Interning Symbols tells you:

   When the Lisp reader encounters a symbol, it reads
   all the characters of the name....  If a symbol
   with the desired name is found, the reader uses
   that symbol.  If the obarray does not contain a
   symbol with that name, the reader makes a new
   symbol and adds it to the obarray.

On the other hand, it is possible, but unusual, for
a symbol to be unnamed, aka "uninterned".  But you
need not worry about this unusual situation when
starting to learn about Lisp.  You should instead
think of a symbol as a Lisp object that _has a name_
(and possibly other properties).  (Node Creating
and Interning Symbols also tells you about
uninterned symbols.)

___

Yes, _read the manual_.  You read a bit about symbols
-- good.  Read that bit again.  And again.  It isn't
rocket science, and lots of effort has gone into
trying to explain things.  Read an introduction to
any Lisp - all Lisps have symbols.  And try the
functions involving symbols.

Read the Intro to Elisp manual that comes with Emacs.
___

[*] More precisely, there cannot be two symbols with
the same name _in the same obarray_.  An obarray is
essentially a symbol namespace.  Each obarray can
have a symbol with any name (that's unique in that
obarray).  Multiple obarrays can each have a symbol
`foo', that is, a symbol named "foo".  But those are
separate, independent symbols, not the same Lisp
object.

Re: [External] : Re: Calling a function with undefined symbol

[Emanuel Berg]

Drew Adams wrote:

> A symbol is indeed a Lisp object - an atom that has a name
> and possibly other properties (value as a variable, function
> definition,...).

Lisp object, atom, property, value, variable, function ... it
never ends ...

Glossary pretty please ...

-- 
underground experts united
https://dataswamp.org/~incal

RE: [External] : Re: Calling a function with undefined symbol

[Drew Adams]

> > A symbol is indeed a Lisp object - an atom that has a name
> > and possibly other properties (value as a variable, function
> > definition,...).
> 
> Lisp object, atom, property, value, variable,
> function ... it never ends ...

It typically ends with nil.

> Glossary pretty please ...

+1.

1. Yes, the Elisp manual could use a
glossary.  (The Emacs manual has one.)

2. Perhaps your best bet is to use `i' in
the Elisp manual.  E.g., `i symbol' takes
you directly to node `Symbols', which tells
you about symbols.

3. It can also help, IMO, to consult CLTL2.
The things we've been discussing here are
mostly common to other Lisps.  (Obviously
things such as buffers and windows are not.
Not so obviously, obarrays are not - CL
uses CL packages instead.) 

https://www.cs.cmu.edu/Groups/AI/html/cltl/cltl2.html

E.g., Symbols:
https://www.cs.cmu.edu/Groups/AI/html/cltl/clm/node27.html

4. As far as glossaries in Info manuals go,
library Info+ can help a bit by fontifying 
and linking glossary words from the text:
options `Info-fontify-glossary-words' and
`Info-link-glossary-words'.  (This includes
links among words within the glossary.)

https://www.emacswiki.org/emacs/InfoPlus#GlossaryEnhancements

But again, the Elisp manual has no glossary.
Nevertheless, Emacs manual glossary entries
(e.g. `list', `variable', `expression') are
linked from the Elisp manual.

Re: [External] : Re: Calling a function with undefined symbol

[tomas]

[-- Attachment #1: Type: text/plain, Size: 663 bytes --]

On Mon, Oct 31, 2022 at 06:10:40PM +0100, Emanuel Berg wrote:
> Drew Adams wrote:
> 
> > A symbol is indeed a Lisp object - an atom that has a name
> > and possibly other properties (value as a variable, function
> > definition,...).
> 
> Lisp object, atom, property, value, variable, function ... it
> never ends ...
> 
> Glossary pretty please ...

See the "Glossary" section in your old and trusty Emacs manual. See
also the "Glossary" entry in the Emacs Wiki [1] which can be found
by feeding your favourite search engine (which hopefully ain't Google:
don't feed that one).

Cheers

[1] https://www.emacswiki.org/emacs/Glossary

-- 
t

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

RE: [External] : Re: Calling a function with undefined symbol

[Drew Adams]

> See also the "Glossary" entry in the Emacs Wiki [1]
> [1] https://www.emacswiki.org/emacs/Glossary

Yes, I meant to also mention the glossary on the wiki.  Thx.

Re: [External] : Re: Calling a function with undefined symbol

[tomas]

[-- Attachment #1: Type: text/plain, Size: 403 bytes --]

On Tue, Nov 01, 2022 at 03:58:57PM +0000, Drew Adams wrote:
> > See also the "Glossary" entry in the Emacs Wiki [1]
> > [1] https://www.emacswiki.org/emacs/Glossary
> 
> Yes, I meant to also mention the glossary on the wiki.  Thx.

But you are right: a reference in the Elisp manual pointing to the
Glossary in the Emacs manual seems like a good idea.

What do people think?

Cheers
-- 
t

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

RE: [External] : Re: Calling a function with undefined symbol

[Drew Adams]

> But you are right: a reference in the Elisp
> manual pointing to the Glossary in the Emacs
> manual seems like a good idea.

I didn't add a reference in the Elisp manual
point to the glossary in the Emacs manual.

(I didn't modify any manuals.  My code just
works with the existing Emacs info files.)

I added the possibility (via user option) of
having occurrences of terms in the Elisp
manual (first occurrence in a topic, for each)
be linked to their entries in the Emacs manual.
___

Actually, there's a user option that lists
the manuals with glossaries to use, in order,
to link to.  By default, only the Emacs and
the Semantics manuals are in that list, since
in the Emacs releases I have those are the
only manuals that have glossaries.

Re: Calling a function with undefined symbol

[Michael Heerdegen]

Stefan Monnier via Users list for the GNU Emacs text editor
<help-gnu-emacs@gnu.org> writes:

> Nitpick:
>
>     'go
>
> is an *expression*, not a symbol.
> That expression returns a symbol, the symbol:
>
>     go

Here is a nice exercise about that:

Say the variable `x` is bound to

  '''''''''''''''''''''''''''''''''...

i.e. a value that is a quoted thing where the thing is the same value
again (such a value exists in Elisp!) - what would evaluating

  (eq 'x x)

and

  (equal 'x x)

yield, and why?


Michael.

Re: Calling a function with undefined symbol

[Sam Steingold]

> * Michael Heerdegen <zvpunry_urreqrtra@jro.qr> [2022-11-02 13:14:55 +0100]:
>
> Say the variable `x` is bound to
>
>   '''''''''''''''''''''''''''''''''...

i.e.,

(defconst x #1=(quote #1#))

right?
NB: (setq print-circle t) before attempting the above!

> i.e. a value that is a quoted thing where the thing is the same value
> again (such a value exists in Elisp!) - what would evaluating
>
>   (eq 'x x)
>   (equal 'x x)

nil: you are comparing a symbol 'x' with a circular list
(quote (quote ...))

what you probably meant was

(eq x (eval x))
(equal x (eval x))

which both evaluate to t because eval just scrolls the circular list
once - to its original starting point

-- 
Sam Steingold (https://aphar.dreamwidth.org/) on darwin Ns 10.3.2113
https://lastingimpactpsychology.com https://steingoldpsychology.com
https://www.peaceandtolerance.org/ https://www.memritv.org
When C++ is your hammer, everything looks like a thumb.

Re: Calling a function with undefined symbol

[Michael Heerdegen]

Sam Steingold <sds@gnu.org> writes:

> >   '''''''''''''''''''''''''''''''''...
>
> i.e.,
>
> (defconst x #1=(quote #1#))
>
> right?

Yep.  Or #1='#1#.  What strange things we can have in Elisp!

> > i.e. a value that is a quoted thing where the thing is the same value
> > again (such a value exists in Elisp!) - what would evaluating
> >
> >   (eq 'x x)
> >   (equal 'x x)
>
> nil: you are comparing a symbol 'x' with a circular list
> (quote (quote ...))
>
> what you probably meant was [...]

No, I really meant what I wrote.  It was intended as that simple
question.

Symbols are values in Elisp, expressions can eval to symbols.  That's
what Stefan was talking about.

The exercise's goal was to recognize that `'x` evals to a symbol, which
is something entirely different than adding another quote to that
strange `'''....` value (which is not a symbol in any case).  If you
have understood that the first value is a symbol, it's easy to give the
intended answers: two times "no".

Michael.

Re: Calling a function with undefined symbol

[Emanuel Berg]

Michael Heerdegen wrote:

> No, I really meant what I wrote. It was intended as that
> simple question.

Uhm, can we have the entire riddle in Elisp please? :)

-- 
underground experts united
https://dataswamp.org/~incal

Re: Calling a function with undefined symbol

[Emanuel Berg]

Michael Heerdegen wrote:

> The exercise's goal was to recognize that `'x` evals to
> a symbol, which is something entirely different than adding
> another quote to that strange `'''....` value (which is not
> a symbol in any case). If you have understood that the first
> value is a symbol, it's easy to give the intended answers:
> two times "no".

Yes, but why do you need the "strange value" for that?

(setq one 1)

(eq    'one one) ; nil
(equal 'one one) ; nil

?

-- 
underground experts united
https://dataswamp.org/~incal

Re: Calling a function with undefined symbol

[Michael Heerdegen]

Emanuel Berg <incal@dataswamp.org> writes:

> Yes, but why do you need the "strange value" for that?

I chose it because you get an `equal' value when you quote the value (to
suggest a wrong track to an answer), and

> (setq one 1)
>
> (eq    'one one) ; nil
> (equal 'one one) ; nil

`1` doesn't have this characteristic: `1` and `'1` are not `equal'
values.


Ok - so here is part two of the exercise to check whether you have
understood quoting.  Is the `length' of this strange value `''''''...`

  (a) 0   or
  (b) 1   or
  (c) 2   or
  (d) infinite/undefined ?

Michael.

Re: Calling a function with undefined symbol

[Emanuel Berg]

Michael Heerdegen wrote:

>> Yes, but why do you need the "strange value" for that?
>
> I chose it because you get an `equal' value when you quote
> the value (to suggest a wrong track to an answer), and

You do?

(setq print-circle t)

(setq x #1=(quote #1#))

(eq    'x x) ; nil
(equal 'x x) ; nil

?

> Ok - so here is part two of the exercise to check whether
> you have understood quoting. Is the `length' of this strange
> value `''''''...`
>
>   (a) 0   or
>   (b) 1   or
>   (c) 2   or
>   (d) infinite/undefined ?

The length is 2 (it's a list) because `quote' does not
evaluate anything, not even itself, however this pattern is
repeated at the second element so there it is either cyclic or
indefinite,

  (quote (quote (quote ... )))

So the length is 2, but the depth is infinite. IOW: It's a trap!

-- 
underground experts united
https://dataswamp.org/~incal

Re: Calling a function with undefined symbol

[Michael Heerdegen]

Emanuel Berg <incal@dataswamp.org> writes:

> >> Yes, but why do you need the "strange value" for that?

Let's call the strange `'''''''''...` values "quote-quine" for now so
that it's easier to talk about them (better than saying "the strange
value" all the time).  I said value_s_ here because there is no unique
quote-quine in the `eq' sense.

> > I chose it because you get an `equal' value when you quote
> > the value (to suggest a wrong track to an answer), and
>
> You do?
>
> (setq print-circle t)
>
> (setq x #1=(quote #1#))
>
> (eq    'x x) ; nil
> (equal 'x x) ; nil
>
> ?

I don't see any place in your code where a quote-quine is quoted.  You
correctly assign a quote-quine to a variable.  But then you only quote
the symbol, not the value, and compare the quote-quine with the symbol
you had bound it to.

IOW: It's a trap!

So how could a correct `equal'ity test of a quote-quine and that
quote-quine quoted be achieved?  This is exercise number 3.


> So the length is 2, but the depth is infinite. IOW: It's a trap!

Yes, this is the correct solution.  Congratulations!

Michael.

Re: Calling a function with undefined symbol

[Michael Heerdegen]

Emanuel Berg <incal@dataswamp.org> writes:

> [...] because `quote' does not evaluate anything, not even
> itself [...]

I'm not so happy with this sentence, however.  Not sure what you
intended to say.

Because, note that quotes (or quoted values, to be more precise) are
_not_ self-quoting: evaluating a quoted value normally yields something
different (i.e., the value, but not the quoted value).

In that point `quote' is different from how `lambda' forms behave in
dynamically binding Elisp ("self quoting"):

  (setq x '(lambda () 1)) ;; x -> (lambda nil 1)
  (eval x nil)            ==> (lambda nil 1)
  (equal x (eval x nil))  ==> t

but OTOH

  (setq x ''nil)         ;; x -> 'nil
  (eval x nil)           ==> nil, not 'nil
  (equal x (eval x nil)) ==> nil

It is important to differentiate between forms and the results of
evaluation of forms the make clear what one is meaning.


Michael.

Re: Calling a function with undefined symbol

[Emanuel Berg]

Michael Heerdegen wrote:

>> because `quote' does not evaluate anything, not even
>> itself
>
> I'm not so happy with this sentence, however. Not sure what
> you intended to say.

If you evaluate `quote' you get the argument but it isn't
evaluated, and this doesn't change if the argument is
another quote.

              (+ 1 2 3)   ; 6
       (quote (+ 1 2 3))  ; not evaluated into 6, instead it's a list
(quote (quote (+ 1 2 3))) ; now it's a list with quote as the first element

And so on.

-- 
underground experts united
https://dataswamp.org/~incal

Re: Calling a function with undefined symbol

[Emanuel Berg]

Michael Heerdegen wrote:

>> (setq print-circle t)
>>
>> (setq x #1=(quote #1#))
>>
>> (eq    'x x) ; nil
>> (equal 'x x) ; nil
>>
>> ?
>
> I don't see any place in your code where a quote-quine is
> quoted. You correctly assign a quote-quine to a variable.
> But then you only quote the symbol, not the value, and
> compare the quote-quine with the symbol you had bound it to.
>
> IOW: It's a trap!
>
> So how could a correct `equal'ity test of a quote-quine and
> that quote-quine quoted be achieved?

Beats me!

Because quoting it will just add another level/layer of the
same, and in the same way, to something that is
already infinite.

(setq print-circle t)

(setq x #1=(quote #1#))

(equal x #1=(quote #1#)) ; t

(equal x (quote #1=(quote #1#))) ; t

(equal x (quote (quote #1=(quote #1#)))) ; and so on

So I give my tongue to the cat then?

-- 
underground experts united
https://dataswamp.org/~incal

Re: Calling a function with undefined symbol

[Michael Heerdegen]

Emanuel Berg <incal@dataswamp.org> writes:

> (setq print-circle t)
>
> (setq x #1=(quote #1#))
>
> (equal x #1=(quote #1#)) ; t
>
> (equal x (quote #1=(quote #1#))) ; t
>
> (equal x (quote (quote #1=(quote #1#)))) ; and so on

You do it... but indirectly.  I wanted to add a quote to the value
already generated.  Like this:

  (equal #1='#1# ''#1#) ==> t

or

  (let ((x #1='#1#))
    (equal x `',x)) ==> t

The backquote-quote-unquote combination is a nicely short way of quoting
an existing value (instead of an expression like a symbol).

But the conclusion is the same as you described.

Michael.

Re: Calling a function with undefined symbol

[Emanuel Berg]

Michael Heerdegen wrote:

> You do it... but indirectly. I wanted to add a quote to the
> value already generated. Like this:
>
>   (equal #1='#1# ''#1#) ==> t
>
> or
>
>   (let ((x #1='#1#))
>     (equal x `',x)) ==> t
>
> The backquote-quote-unquote combination is a nicely short
> way of quoting an existing value (instead of an expression
> like a symbol).
>
> But the conclusion is the same as you described.

Can't you put these and other such questions you may know of
into a document and others can add theirs ... don't know if
I know any?

-- 
underground experts united
https://dataswamp.org/~incal

Re: Calling a function with undefined symbol

[Michael Heerdegen]

Emanuel Berg <incal@dataswamp.org> writes:

> Can't you put these and other such questions you may know of

These were just improvised for fun...

> into a document and others can add theirs ... don't know if I know
> any?

In my objective point of view the canonical person for the task
is...Emanuel Berg! (not kidding)

Michael.

RE: [External] : Re: Calling a function with undefined symbol

[Drew Adams]

> Emanuel Berg <incal@dataswamp.org> writes:
> 
> > Can't you put these and other such questions you may know of
> 
> These were just improvised for fun...
> 
> > into a document and others can add theirs ... don't know if I know
> > any?
> 
> In my objective point of view the canonical person for the task
> is...Emanuel Berg! (not kidding)

Open encouragement/invitation for The Canonical Berg to do it!

(Anyone can create pages on Emacs Wiki, for instance.)

Re: Calling a function with undefined symbol

[Emanuel Berg]

Michael Heerdegen wrote:

>> into a document and others can add theirs ... don't know if
>> I know any?
>
> In my objective point of view the canonical person for the
> task is...Emanuel Berg! (not kidding)

Okay, I'll get right to it!

Hey, wait ... nice try :)

-- 
underground experts united
https://dataswamp.org/~incal

Re: Calling a function with undefined symbol

[Michael Heerdegen]

Emanuel Berg <incal@dataswamp.org> writes:

> > In my objective point of view the canonical person for the
> > task is...Emanuel Berg! (not kidding)
>
> Okay, I'll get right to it!

Good.  You may add

(d) What's the return value of evaluating the expression

  (funcall
   (lambda (x) (funcall x x))
   (lambda (x) (funcall x x)))


> Hey, wait ... nice try :)

Eh - dito.


Michael.

Re: Calling a function with undefined symbol

[tomas]

[-- Attachment #1: Type: text/plain, Size: 634 bytes --]

On Tue, Nov 08, 2022 at 04:52:38PM +0100, Michael Heerdegen wrote:
> Emanuel Berg <incal@dataswamp.org> writes:
> 
> > > In my objective point of view the canonical person for the
> > > task is...Emanuel Berg! (not kidding)
> >
> > Okay, I'll get right to it!
> 
> Good.  You may add
> 
> (d) What's the return value of evaluating the expression
> 
>   (funcall
>    (lambda (x) (funcall x x))
>    (lambda (x) (funcall x x)))

Now would that be more fun if Elisp did tail call optimization?

Let me try in Guile:

  scheme@(guile-user)> ((lambda (x) (x x)) (lambda (x) (x x)))

Yes, it is :)

Cheers
-- 
t

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

Re: Calling a function with undefined symbol

[Michael Heerdegen]

<tomas@tuxteam.de> writes:

> Now would that be more fun if Elisp did tail call optimization?
>
> Let me try in Guile:
>
>   scheme@(guile-user)> ((lambda (x) (x x)) (lambda (x) (x x)))
>
> Yes, it is :)

Impressive - it just produces a bit of heat over time, but it doesn't
consume memory.

Michael.

Re: Calling a function with undefined symbol

[tomas]

[-- Attachment #1: Type: text/plain, Size: 533 bytes --]

On Tue, Nov 08, 2022 at 08:57:59PM +0100, Michael Heerdegen wrote:
> <tomas@tuxteam.de> writes:
> 
> > Now would that be more fun if Elisp did tail call optimization?
> >
> > Let me try in Guile:
> >
> >   scheme@(guile-user)> ((lambda (x) (x x)) (lambda (x) (x x)))
> >
> > Yes, it is :)
> 
> Impressive - it just produces a bit of heat over time, but it doesn't
> consume memory.

To be more precise, it always consumes the same memory. Perhaps the
bits become somewhat worn out in that place ;-P

Cheers
-- 
t

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

Re: Calling a function with undefined symbol

[Sam Steingold]

> * Michael Heerdegen <zvpunry_urreqrtra@jro.qr> [2022-11-08 16:52:38 +0100]:
>
> (d) What's the return value of evaluating the expression
>
>   (funcall
>    (lambda (x) (funcall x x))
>    (lambda (x) (funcall x x)))

or this:

((lambda (x) `(,x ',x)) '(lambda (x) `(,x ',x)))


-- 
Sam Steingold (https://aphar.dreamwidth.org/) on darwin Ns 10.3.2113
https://lastingimpactpsychology.com https://steingoldpsychology.com
http://think-israel.org https://honestreporting.com https://www.memritv.org
Telling a programmer there's already a library to do X is like telling a songwriter there's already a song about love.

Re: Calling a function with undefined symbol

[Michael Heerdegen]

Sam Steingold <sds@gnu.org> writes:

> or this:
>
> ((lambda (x) `(,x ',x)) '(lambda (x) `(,x ',x)))

Then I would say

  (#1=(lambda (x) '(#1# '#1#)) #1#)

to get rid of duplicated lambdas and backquotes in the code (isn't it
much better readable now?).

Michael.

Re: Calling a function with undefined symbol

[Sam Steingold]

> * Michael Heerdegen <zvpunry_urreqrtra@jro.qr> [2022-11-09 00:53:33 +0100]:
>
> Sam Steingold <sds@gnu.org> writes:
>
>> ((lambda (x) `(,x ',x)) (lambda (x) `(,x ',x)))
>
>   (#1=(lambda (x) '(#1# '#1#)) #1#)

Nah, (#1=(lambda (x) `(,x ',x)) #1#)

> (isn't it much better readable now?).

I would rather avoid circular lists _in code_.
(and your version is _not_ the same as mine)

-- 
Sam Steingold (https://aphar.dreamwidth.org/) on darwin Ns 10.3.2113
https://lastingimpactpsychology.com https://steingoldpsychology.com
https://www.dhimmitude.org https://memri.org https://iris.org.il
Lottery is a tax on statistics ignorants.  MS is a tax on computer-idiots.

Re: Calling a function with undefined symbol

[Michael Heerdegen]

Sam Steingold <sds@gnu.org> writes:

> > * Michael Heerdegen <zvpunry_urreqrtra@jro.qr> [2022-11-09 00:53:33
> > +0100]:
> >
> > Sam Steingold <sds@gnu.org> writes:
> >
> >> ((lambda (x) `(,x ',x)) (lambda (x) `(,x ',x)))
> >
> >   (#1=(lambda (x) '(#1# '#1#)) #1#)
>
> Nah, (#1=(lambda (x) `(,x ',x)) #1#)

Yes, I was wrong two times: the first time when I posted the wrong
counterpart of the quote-quine.  My version without quoting was a
slightly different thing because evaluation didn't stop.  Your version
fits better.

And (lambda (x) '(#1# '#1#)) is nonsense because it ignores the
argument.

But why not ((lambda (x) `(,x ,x)) (lambda (x) `(,x ,x))) - why do you
quote the second list element?  I guess you want to avoid that the
lambda is evaluated and gets a closure, but then you want

  ((lambda (x) `(,x ',x)) '(lambda (x) `(,x ',x)))

                          ^
> I would rather avoid circular lists _in code_.

I had been wondering about the question: If

  ((lambda (x) `(,x ',x)) (lambda (x) `(,x ',x)))

is the version for n=1 arguments, how would a version for n=0 look like?
Then you could either construct (a) an expression that returns the n=0
counterpart, which is easy, or give (b) an expression that is the
counterpart itself.  Then I think you need circular code because the
function does not receive an argument:

  (#1=(lambda () '(#1#)))


Michael.

Re: Calling a function with undefined symbol

[Emanuel Berg]

Michael Heerdegen wrote:

>>> In my objective point of view the canonical person for the
>>> task is...Emanuel Berg! (not kidding)
>>
>> Okay, I'll get right to it!
>
> Good. You may add (d)

OK, I'll do it! But what were (a) to (c) exactly since there
were so much discussion ...

I'll add my own egg boiler question!

When we have say - I don't know how many - we will have Drew
mail the Emacs Wiki maintainer and ask if he can add it
there ...

> What's the return value of evaluating the expression
>
>   (funcall
>    (lambda (x) (funcall x x))
>    (lambda (x) (funcall x x)))

I don't know! :O

What is it?

-- 
underground experts united
https://dataswamp.org/~incal

RE: [External] : Re: Calling a function with undefined symbol

[Drew Adams]

> When we have say - I don't know how many - we will have Drew
> mail the Emacs Wiki maintainer and ask if he can add it
> there ...

Hm.  I've said this several times before, but once more...

_Anyone_ can create and edit pages on Emacs Wiki.

And the home page tells you this and how to do it.

https://www.emacswiki.org/emacs/SiteMap

See the "How to use this site" links on that page.
In particular, "How To Edit":

https://www.emacswiki.org/emacs/HowToEdit,

which has a link to page "Create New Pages":

https://www.emacswiki.org/emacs/CreateNewPages

Try it!  Create a page.  If you don't want it to
stay up, then delete it:

https://www.emacswiki.org/emacs/DeletedPage


And there's a "Sand Box" page, where you can try
out how to write/edit different things, to get
the appearance you want.

https://www.emacswiki.org/emacs/SandBox

___

No need for anyone to email the wiki maintainer
and ask him for anything.  About the only thing
I email him for is if I think there's some kind
of problem with the site, e.g., the server needs
to be rebooted.

> > What's the return value of evaluating the expression
> >   (funcall
> >    (lambda (x) (funcall x x))
> >    (lambda (x) (funcall x x)))
> 
> I don't know! :O
> What is it?

Enter (λ x.xx)(λ x.xx) here:

https://lambdacalc.io/

Re: [External] : Re: Calling a function with undefined symbol

[Emanuel Berg]

Drew Adams wrote:

>>> What's the return value of evaluating the expression
>>>
>>>   (funcall
>>>    (lambda (x) (funcall x x))
>>>    (lambda (x) (funcall x x)))
>> 
>> I don't know! :O
>> What is it?
>
> Enter (λ x.xx)(λ x.xx) here:
>
> https://lambdacalc.io/

?

Why do you say that? Everyone can use that site.

-- 
underground experts united
https://dataswamp.org/~incal