[PD] abstractions which have their own memory

Chris McCormick chris at mccormick.cx
Thu Jul 6 06:35:39 CEST 2006


On Wed, Jul 05, 2006 at 04:51:26PM +0200, Frank Barknecht wrote:
> Hallo,
> Christopher Charles hat gesagt: // Christopher Charles wrote:
> 
> > most of my bagoftricks patches look radically different and i don't
> > save copies of them with only some parameters changed. another bonus
> > for me is ease of use, with clicking pd's save button, i save
> > exactly what i'm working on at the moment, when i load it up later,
> > it sounds the same. only one file to take care of. 
> 
> Of course to each his own, but pressing a different "SAVE" to save the
> patch than to save the current state of the variable in my patch is no
> problem for me. 

Sure, that is a really good way of working, but not everybody wants to
patch like that. In my opinion if puredata had the following features
then it would make it a lot easier to make patches that save their own
state within the main .pd file and still be able to use abstractions
for modularity, which is the way some people like to work.

1. [parentname] which when banged gives the name of the parent patch.

2. [savebang] which bangs each time ctrl-s is hit or "save" is selected
from the menu.

This would enable GOP abstraction units to save their internal data to
the parent patch (or a sub-patch of the parent patch) whenever the patch
is saved.

At the moment you are forced to work in the Frank B way; make lots
of virtual machines that save their state to a file collectively
or individually, and can have multiple states for one "virtual audio
machine" configuration, or the Hans Steiner/Solitude way; make a big
patch which uses data structures to save it's state/score but can't
really use abstractions meaningfully within that framework.

They way I like to work, and obviously some others too, is for each patch
to be a self contained score and instrument, but to have the re-use of
abstractions available. This is currently pretty difficult, so I use
[sssad] to save state (thanks for that!). It would be nice to be able to
build something that allowed for this alternative.

Best,

Chris.

-------------------
chris at mccormick.cx
http://mccormick.cx




More information about the Pd-list mailing list