current goal/milestone

ensure minimum use-case is possible = working on all devices, with raw mapping, with lights/text control.
does not include: scales, splits,

current priorities

(in order)

acheive feature parity, api integration for all devices
add ability to drives leds for eigenharps via osc
add ability to drive push 2 parameters/display via osc
packaging of mec-app

mec-api , known issues

push2 , I call render in process(), it probably should be in a separate low prio thread.
consider libusb initialisation, may be multile libusb devices running

mec-api , planned changes - mec-api

eh led control via osc
push 2 display via osc
better surface mapping
runtime connect/disconnect of devices
multiple instances of devices
allow api to say which devices to start... so can run multiple instances using same mec.json
runtime configuration changes (properties) (mec-kontrol will implement this)
push 2 support, using properties/runtime config (mec-kontrol will implement this)

mec-api , windows support

perhaps check with MGwin build (allows pthread)
rtmidi uses pthread, move to juce?
adapt cmake
soundplane lib
move to use std::thread in mec-app, remove any other pthread dependancy

build issues

brew install issue on mac, as no longer supply universal binary, build libusb?
properly package some of the mec apps (some need manual intervention)
mec-app rpath
mec-vst need to copy dylibs into Frameworks
mec-max need to sort out rpath for external, and package dylibs
consider a top level script to build everything (perhaps options depending on platform being run on... e.g. no point in bela on windows)
remove OSAtomic* errors from eigenfreed/eigend, use std::atomic_* instead

work in progress

scales
splits/combining of surfaces
mec-kontrol

refactor/generalising

voices data, a few places where I do similar things to collect the data
osc generation, create a processor (issue is oscpack, doesnt have a message as such... just a stream, does this add much?... do we need osc output elsewhere?)
reconsider rtmidi vs juce , rtmidi is dependent on pthread, so perhaps juce

improvements

osc/t3d ouput (mec-app) , output periodic framemessage (/t3d/frm)
osc/t3d input, track /t3d/dr, then look for /t3d/frm cancel voices if not received in time
config - device class and instances...

other

document configuration file/setup
improve existing docs

device specific to do

soundplane

upgrade to new touch tracker (when stable/available)

pico

mode keys
windows/threshold/warmup for strip(s)/breath

tau

test key layout due to 'odd shape'
mode keys
windows/threshold/warmup for strip(s)/breath

alpha

windows/threshold/warmup for strip(s)/breath

push2

refactor to allow for different control/navigation modes