On Feb 2, 2012, at 11:53 AM, Peter Brinkmann wrote:

> 
> 
> On Thu, Feb 2, 2012 at 1:52 AM, Rich E <[email protected]> wrote:
> I do think it is important to separate these things into bite size chunks (I 
> think IOhannes mentioned this as well during his LAC talk).  Peter, your blog 
> post talks of creating an API for editing patches (here), and while I look 
> forward to these capabilities, I think this is also a separate job as to the 
> one Miller proposed on this thread, which I see as taking care of the static 
> state in pd.  I don't think I could prioritize these two different jobs, but 
> I'd say multiple instances allows us to definitively crush max, as we'll have 
> a pd vst. :)
> 
> I harbor no hostility to Max, but I agree that getting rid of global state 
> and allowing multiple instances should be the first order of business.  Any 
> ideas?


The way I see the "editing API" is the converse of the pd-gui --> pd 
communications: pd messages.  My approach is to do it bit by bit.  Take a chunk 
of the current pd-gui --> pd communications and refactor into something that 
looks like a pd message (Tcl proc calls can look the small, luckily, the syntax 
is similar in some basic ways).

Ico has recently refactored array moving into a single 'move' command, that's 
in pd-l2ork.  I haven't looked at that yet.  But that's the same idea.

.hc

----------------------------------------------------------------------------

"A cellphone to me is just an opportunity to be irritated wherever you are." - 
Linus Torvalds

_______________________________________________
Pd-dev mailing list
[email protected]
http://lists.puredata.info/listinfo/pd-dev

Reply via email to