[bang(
|
[symbol 1]
|
[print]

... prints only "symbol."

It's illegal to have an ASCII representation of a number stored in a symbol object?

Well, I disagree with that. If it's an abstraction argument that could be either numeric or symbolic, then... the only way is [pdcontrol], right? Because you don't know in advance whether it should be [f $1] or [symbol $1]. (Come on... ftoa() isn't hard.)

hjh

@ddw_music Yes, [pdcontrol] + [route] is the way...

@ddw_music Otherwise [makefilename] can make a symbol of a number...

@ddw_music +1 to [makefilename], but in instances where I also want to stop Pd from truncating my float, I prepend an escaped space to force Pd to treat the string of numbers as a symbol. [f] is happy to convert it back to a number when needed.

It was discussed in github and hopefully will be fixed one day on pd side https://github.com/pure-data/pure-data/pull/987

Thanks all -- I ended up using pdcontrol because the final target is a message box (so in fact it could be either a number or a symbol).

I'm glad to hear it's been discussed. Hope it doesn't disappear into that abyss from which pull requests never return.

hjh

in max we dont have a [symbol] object so we would use for example

[f 3.17]
[tosymbol]

or

[t "3.17"]

both does not to work in pd, and not even [trigger foo] is allowed, so it seems a bit strange that it doesnt work with [symbol] either.

because... what is its purpose then?

what works fine in pd is [expr (3.123456789)] - that is what i also use in max if i want to see more than the truncated 32 bit value.

@Obineg said:

in max we dont have a [symbol] object ...
so it seems a bit strange that it doesnt work with [symbol] either.

because... what is its purpose then?

Off topic, but the punchline I found most recently in Max is that [swap] allows only ints for the left inlet / right outlet

Plenty of inconsistencies to be found in both platforms

hjh

back in max v.4 a bunch of objects such as [cycle] only took type int - be happy when you missed that time.
with the years i replaced most of them with custom abstractions. i dont care for the excus... pardon, reasoning of this behaviour, i want to work with arbitrary data without thinking.

i mean it is funny: in 1990, on the one hand they wanted to avoid the CPU cycles for the type check for the midi objects, but then they did the same for some non midi releated objects,

on the other hand they were always using LONG from day 1, where they should have introduced types of 16 or 8 bit numbers from the beginning on and save fare more from that than from a type check.