[PD] How's Pd limited?
jancsika at yahoo.com
Tue Feb 23 00:17:24 CET 2016
> It's just that the presence of those features makes it much easier not to care, and many users just don't care, and it makes things worse for those of us who have to use that patch elsewhere.
Short story: I'm not going to write the code to implement segmented cords, and
I don't think anyone else is, either. But if someone wants to I'll certainly have
Non-segmented patch cords suggest a flow chart. Segmented patch cords--
when used sensibly-- suggest a circuit board. The difference IMO is nobody in
their right mind would suggest circuit boards are an especially readable and
friendly way to elucidate a program's logic to human beings. That's the starting
premise of every "flow based" language, though, and I've never seen much
evidence to back it up.
All those awful-looking yet functional Pd and Max patches get that way
because of the strength of the approach, IMO-- that is, a physical motion with
visual animated feedback creates the flow of data. It's the drawing of the
patch matters, not the engraving.
For example, imagine that Pd tracked Dan's eye movement on the first
version of his patch. Those objects and wires that he hasn't looked at for
awhile fade more and more into the background, unless he tries to focus
on one of them and then they return. I think that's a pretty decent description
of how we actually create those write-only patches. Throw in whatever the
opposite of MVC is (e.g., where a toggle can appear smack dab in the model),
and you have an environment that's well suited to quick prototyping.
Of course that mental model has a short shelf-life, so there's the separate
issue of how to turn that into a readable patch. There are certainly patterns
to follow there, and ways to minimize the spaghetti in the first place. But I think that potential to draw big ugly lines across the whole damn thing is what drives the speed and elegance of developing in the language. If it weren't then Pd
would be like brainfuck, and none of those spaghetti patches would be able to
deliver any functionality to speak of.
So for those users who don't care about taking their patches (or, hopefully,
small abstractions or subpatches) from
"draw-time ugly-mode" to "presentable-to-other-humans mode", I don't think
segmented cords matter much. For the ones who do, I guess I'd rather look at
a flow chart than a circuit board. But given the choice I'd rather watch little
gremlins carry buckets of water up a hill, or robots shooting lasers at drones.
Or at least see the data "pumping" through all those boring control wires that seem to always obscure the text I'm trying to read...
On Monday, February 22, 2016 5:20 PM, Matt Barber <brbrofsvl at gmail.com> wrote:
Forcing good practice is not something I'm interested in. Every programming language can be abused horribly (they even have a prize for best/worst abuse of C -- look through some of these http://www.ioccc.org/years.html ).
My point was not that avoiding segmented and hideable patch cords fixes these problems. It's just that the presence of those features makes it much easier not to care, and many users just don't care, and it makes things worse for those of us who have to use that patch elsewhere.
We did have a long list discussion about best practices, actually, collected here:
There are other style guides too.
On Mon, Feb 22, 2016 at 4:29 PM, Dan Wilcox <danomatika at gmail.com> wrote:
2016-02-22 17:25 GMT-03:00 Matt Barber <brbrofsvl at gmail.com>:
I've said this before, but I think there are very good reasons not to ever include segmented patch cords (although hideable patch cords would be even worse). These two features are responsible for some of the very worst patching habits in Max/MSP. Have you ever been called on to run someone's patch, and you need to tweak something for your specific audio setup or fix a bug or whatever, and when you open it you get something that looks like this (one of the first "max patch" results on google image search):
I agree, and I laugh when people say, this is hard to understand in Max, because of all the cords, I can't imagine how ugly it'd be in Pd.
The solution is the same in both environments: good use of encapsulation via subpatches & judicious use of send/recvs when necessary.
Example from robotcowboy:
* one of my first performance patches: https://www.flickr.com/photos/danomatika/25082084442/in/datetaken-public/* and the second version using subpatches & send/recvs: https://www.flickr.com/photos/danomatika/24573539133/in/datetaken-public/
This approach led to using GOP and modularizing things between separate patches & a main control patch: https://www.flickr.com/photos/danomatika/25107115651/in/datetaken-public/
To segment or not to segment is moot, you can create both well designed as well as spaghetti patches in either environment just as you can create well-written or spaghetti code in any textual language. I agree that the environments are not at fault here.
Pd-list at lists.iem.at mailing list
UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Pd-list