So I attended and presented at the PAW, https://faust.grame.fr/paw/ today and
I listened carefully to most of the talks and I have a few thoughts following it.

1) There’s a lot of activity in the embedded space and we should pay
attention to. We have already a presence in Bela, and we need to look
at other platforms:

- SAM has Faust programs running on a dedicated Sharc device that
has an audio and MIDI shield. However, it’s bare metal and the toolchain
does not look mature. Not sure if it is possible to target it. I was going to
ask about it at the afternoon workshop, but it was a bit shambolic and
then I had to go and do my own workshop.

-  OWL is another bare metal platform and it has a C++ API for 
development. However with no filesystem and other things we are used
to, I don’t think it is possible to think of porting Csound to it without major
surgery.

- MOD runs on Linux and it’s effectively a LV2 host. This looks very
achievable. We need to design a LV2 plugin frontend, but it should
be OK (provided they have libsndfile in their image, just thought about
it now).

2) The SOUL thing doesn’t look too bad, I think it would be interesting
to see what comes out of it. For us, it might be another way of running
an embedded JIT compiler to do custom opcodes (in addition to Faust).
I had a chat with Jules and lunchtime and it all seems reasonable.

3) Faust seems to be supporting absolutely everything, which is
a plus for them. On the other hand, the fact that multirate processing
is still not available limits it (no spectral processing for instance). There
was a nice recursive FFT program by Julius Smyth shown in the talk
today, but it can only run on a sample-by-sample basis, which of
course kills it (surprisingly they did not consider making a sliding FFT
which would be faster in this case).

WebAudio Csound seems to have gone down well, we had good 
attendance at the afternoon workshop where I outlined how it worked
and showed some examples. We did not have time for more hands-on,
but the material is up on https://vlazzarini.github.io/paw for anyone
wanting to look at it.

========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 708693

A very late reply, but thanks very much Victor for writing this up.
Besides core Csound and Blue dev, I'm looking forward to getting to a
number of Web Csound based projects for 2019.  (I was thinking of a
CsoundXR project for using Csound with extended reality, i.e., WebXR).

Speaking of embedded, I was looking at Rust the other day and thought
it might be nice to experiment with FFI and add to the
CsoundAPI_examples project.  Rust advertises efficiency and working
well with embedded, so it could be another path to promote if
experiments prove to be interesting.

On Sat, Dec 1, 2018 at 12:56 PM Victor Lazzarini  wrote:
>
> So I attended and presented at the PAW, https://faust.grame.fr/paw/ today and
> I listened carefully to most of the talks and I have a few thoughts following it.
>
> 1) There’s a lot of activity in the embedded space and we should pay
> attention to. We have already a presence in Bela, and we need to look
> at other platforms:
>
> - SAM has Faust programs running on a dedicated Sharc device that
> has an audio and MIDI shield. However, it’s bare metal and the toolchain
> does not look mature. Not sure if it is possible to target it. I was going to
> ask about it at the afternoon workshop, but it was a bit shambolic and
> then I had to go and do my own workshop.
>
> -  OWL is another bare metal platform and it has a C++ API for
> development. However with no filesystem and other things we are used
> to, I don’t think it is possible to think of porting Csound to it without major
> surgery.
>
> - MOD runs on Linux and it’s effectively a LV2 host. This looks very
> achievable. We need to design a LV2 plugin frontend, but it should
> be OK (provided they have libsndfile in their image, just thought about
> it now).
>
> 2) The SOUL thing doesn’t look too bad, I think it would be interesting
> to see what comes out of it. For us, it might be another way of running
> an embedded JIT compiler to do custom opcodes (in addition to Faust).
> I had a chat with Jules and lunchtime and it all seems reasonable.
>
> 3) Faust seems to be supporting absolutely everything, which is
> a plus for them. On the other hand, the fact that multirate processing
> is still not available limits it (no spectral processing for instance). There
> was a nice recursive FFT program by Julius Smyth shown in the talk
> today, but it can only run on a sample-by-sample basis, which of
> course kills it (surprisingly they did not consider making a sliding FFT
> which would be faster in this case).
>
> WebAudio Csound seems to have gone down well, we had good
> attendance at the afternoon workshop where I outlined how it worked
> and showed some examples. We did not have time for more hands-on,
> but the material is up on https://vlazzarini.github.io/paw for anyone
> wanting to look at it.
>
> ========================
> Prof. Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy,
> Maynooth University,
> Maynooth, Co Kildare, Ireland
> Tel: 00 353 7086936
> Fax: 00 353 1 7086952