[struct ex1 float y float i glist foo pd-bar] <-- glist field "foo"
using [pd bar] as a template
I think this would be a good idea. specially if you can change the patch
name, and the structure "jumps" as well between subpatches.
two other things I just remember now:
- click/move/change reports for scalars are quite good, because it's very
easy to know which scalar is being edited/active. but they should also be
integrated in arrays.
- as great as the improvements might be, the work will be only partial if
the gui doesn't get improved. since data structures depend 100%
graphically on tcl/tk graphics, they're very slow to work with, if you're
doing something complex. in order for data structures to be more widely
used on realtime patching/editing, the tcl/tk graphical code should be
more efficient.
- give a nameable string to each data-s, which keeps during sessions. this
way is easier to get a determinate data-s in the middle of a big window.
for now it's necessary to go through all the scalars in the window, and
match any inner parameters (an index, a symbol, etc.)
Data structures can be a very powerful mean to graphically
display/interact with the patch (to perform, display, compose, ...). But
without some improvements they'll still be a hidden part of pd. And afaik,
this is the only major feature that separates pd from max - but maybe I'm
wrong?