[CM] Re: [PD] Snd as a pd external documentation.

Hans-Christoph Steiner hans@eds.org
Fri, 10 Feb 2006 00:14:49 -0500


On Feb 9, 2006, at 4:26 PM, Kjetil S. Matheussen wrote:

>
> Okey. But I'm not sure its a good idea. Snd objects are not  
> deterministic. If you try to write an object in a different way  
> (external/subpatch/pyext/etc.) that is supposed to to the same, it  
> will
> 99% for sure behave slightly different.
>
> The reason is that data-handling in the snd external are threaded and  
> is run asynchronus to pd, and that all the dsp-handling is performed  
> at once (once per cycle) for all snd objects, because of performance  
> reasons. (dsp using the snd external is extremely efficient)
>
> I'm afraid hiding these facts from the user can create confusing  
> situations...

yes indeed, I had no idea it was structured like that.

> Regarding the system path settings, aren't they available somewhere  
> anyway?


> And whats a "newly developing namespace"?

The first steps towards implementing a namespace (like java's import  
java.swing.*, for example) are in place.  You can use objects of the  
same name with a namespace prefix, like [cyclone/prepend],  
[iemlib/prepend], [ggee/prepend], etc.  Then you could [import  
cyclone], and [prepend] would then be cyclone's prepend.  The next step  
is to make patch-level namespace, currently there is only a global  
namespace.

.hc

>
>
>
> On Thu, 9 Feb 2006, Hans-Christoph Steiner wrote:
>
>>
>> This would allow you to write bonafide Pd objects in snd.  What  
>> language they are written in would be transparent to the user.  They  
>> would behave just like objects.  They could be bundled into  
>> libraries, they would work with the system path settings, they would  
>> work in the newly developing namespace, they could have their own  
>> help patches, etc. etc.
>>
>> .hc
>>
>> On Feb 9, 2006, at 12:10 AM, Kjetil S. Matheussen wrote:
>>
>>> Hmm, I don't see the point. Having those extra four letters ("snd ")  
>>> shows that it is an snd object, and not an external or a patch.  
>>> Whats the point of hiding that information?
>>> On Wed, 8 Feb 2006, Hans-Christoph Steiner wrote:
>>>> It would be great if this was set up to use Pd's new generic loader  
>>>> functionality.  Then you could write Pd objects directly in  
>>>> Guile/snd without needing the [snd] object. It looks like you are  
>>>> most of the way there.
>>>> There is currently a Common Language Runtime (C#, etc) API in the  
>>>> works that is being ported to the new loader also.
>>>> Basically, you just write a little loader program that tells Pd how  
>>>> to load an object that is written in snd.
>>>> .hc
>>>> On Feb 8, 2006, at 5:52 PM, Kjetil Svalastog Matheussen wrote:
>>>>> Hi, this mail is to inform that Bill Schottstaedt just put up  
>>>>> documentation for the snd pd external in the snd-manual:
>>>>> http://ccrma.stanford.edu/software/snd/snd/grfsnd.html#sndwithpd
>>>>> The documentation also also contains a screen-shot of a pretty neat
>>>>> stochastic pd patch made by Bill Sack, which use the snd external.
>>>>> If you know (and like) lisp, pd and snd, you should check this  
>>>>> out...
>>>>> _______________________________________________
>>>>> PD-list@iem.at mailing list
>>>>> UNSUBSCRIBE and account-management ->  
>>>>> http://lists.puredata.info/listinfo/pd-list
>>>> ____________________________________________________________________ 
>>>> __ ______
>>>>                    There is no way to peace, peace is the way.
>>>> 						        				-A.J. Muste
>>
>> ______________________________________________________________________ 
>> ______
>>
>> If you are not part of the solution, you are part of the problem.
>>                                                              -  
>> Eldridge Cleaver
>>
>>
>>

________________________________________________________________________ 
____

"Computer science is no more related to the computer than astronomy is  
related to the telescope."
                                                           -Edsger  
Dykstra