I once heard miller Puckette in a lecture say that its possible for pure data to write its own patch? Does this mean you can design a patch and save the code as a script and feed that to pure data when you need it? Basically I wasn't sure what he meant entirely and was hoping someone had heard of this ?
Thank you very much for reading?

yup, see the textfile object and open up a patch with a text editor and you'll see that all you need to do to write a patch using pd is to put the lines in the patch into the textfile one by one. the protocol for making pure data messages is called fudi and it allows you to create scripts to create pure data patches in pd or another language. this can also be used for dynamic patching in pd.
http://en.wikipedia.org/wiki/FUDI
also see pd-msg folder in the manuals

Interesting thank you, what are some of the main advantages to this?

Dynamic coding/patching is a very interesting and strong concept in pd (and in any language actually). In the case of pd, it allows you to dynamically create new patches or part of patches depending on your needs. This way you could for instance, in the case of a synth, create new voice abstractions at run-time when more voices are required instead of having a hard-coded and pre-patched limit. In the case of a neural network it would allow you to create neural nodes on the fly.

I'm sure the are plenty more examples where dynamic patching can be very usefull.

Awsome, Thank you very much

I noticed your "pitch tracking" track on sound cloud, thought that was really interesting. Was this through pure data? If it was can the objects you were using identify the voice and play it back at a particular quantised pitch, like a auto correction type function ? Im interested as Im in the middle of experimenting myself with this sort of patch
Thanks again

I used the beautiful [helmholtz~] object to do the pitch tracking (search the forum for it). Here I used it to do analog style continuous tracking, but there would be no problem to quantize this against a given scale. Just look for the object, it also has a good help file explaining how things work.

Thank you

With the dynamic patching I think Im getting it wrong so far. I started off looking at this page http://jeraman.info/2009/03/22/how-to-use-pure-data-as-a-api/ but even on step one when I copied the -nogui commands my terminal on OSX didn't recognise the command I was wondering if anyone new any good sources online where It gives some guidance as to executing affective dynamic patching? I've become familiar with the concept of opening a pd file in a textfile its just I haven't found a way to use it afterwards in action, Perhaps its a matter of continuing to search the forum?