[PD] what distinguishes a numeric symbol as an argument?

[Frank Barknecht]

Hallo,
Matteo Sisti Sette hat gesagt: // Matteo Sisti Sette wrote:

> However, I can't see any difference between the output of the message box 
> (*) in (4) and in (1-3).
> If I connect that message box to a [print] in any of the four examples 
> above, the output is always "label 43".

[print] is very bad for showing these kinds of things. 

> However, this does not explain the following:
> 
> * Isn't the output of [makefilename] in example (4) just the same as [symbol 
> 43( of example 3?? 

It isn't: [makefilename] always generates real symbols and it's the
only (usual) way to really numeric symbols. There was a thread on
pd-dev about this some time ago. The 43 in a message box [symbol 43(
is not a real numeric symbol. Actually "symbol 43" is kind of an
illegal message in Pd, as a proper symbol-message is built from the
word "symbol" and another "word", not a number. So to build a proper
word "43" first you need to generate a 43-symbol using [makefilename].

> Then why doesn't example 3 work as well???
> 
> Also, consider this:
> 
> (5)
> [label 43(
> |
> [unpack s f]
>
> and this
> 
> (6)
> [symbol 43(
> |
> [label $1(               (**)
> |
> [unpack s f]
> 
> In example (5), the second outlet of unpack outputs float 43, which is 
> (somewhat) coherent with the fact that example (1) didn't succeed to set the 
> canvas label
> However, in example (6) unpack gives the error "unpack: type mismatch"... so 
> I guess that the "$1" (i.e. the "43") in (**) is seen as a symbol, not a 
> float....... 

It's neither a symbol nor a float: To be a float, it would need to
have a selector of "float" or no selector at all, to be a symbol, the
data part of the message should be a symbol, but it is a float (43) so
your message is not valid. 

> Oh shit!!!!!!!
> 
> Now I see that the following:
> [symbol 43(
> |
> [print]
> 
> just outputs "symbol "

Yes, [print] gets confused about the invalid message.

> while
> [43(
> |
> [makesymbol %d]
> |
> [print]
> 
> outputs "symbol 43" as expected....

However this "43" now is indeed a symbol and you cannot input it into
a numberbox anymore. Try this: 

 [43(
 |
 [makefilename %d]
 |
 [$1(
 |
 [+ 10]

You get the error message: 

error: +: no method for '43'

which indicates, that "43" now is a symbol-selector! (Let that sink
for a while ...) Numeric symbol selectors aren't used by many objects.
[select] can handle them, you can label GUI elements with them, you
can use them as targets for [send] through the second inlet, but you
cannot build a receiver for them: [r 43] is not valid, as in this case
"43" is still a float.

> isn't it weird??????

Ah, the fun of Pd's message system! But once you get to grips with it
you may see, that it's limited, but actually rather consistent. It'd
recommend an afternoon of list-archive reading, as this confuses many
people.

Ciao
-- 
 Frank Barknecht                 _ ______footils.org_ __goto10.org__