1st Conreality Hackathon report

[Arto Bendiken]

I am pleased to report that the inaugural Conreality Hackathon took
place this past weekend (November 7-8) at the Gogulski Batcave in
Bratislava.

Attending were @mikegogulski, @bendiken, and @DanKomorny, with the
event presided over by his unhandled feline highness Charlie.

We made good progress validating and integrating the hardware for the
initial POC project [1], based on the two Raspberry Pi main boards
provided by Mike. The primary outcomes of this hackathon were as
follows:

1. Validated a native build of the Conreality platform with an
up-to-date OCaml and OPAM toolchain on a Raspberry Pi running Debian
Jessie [Mike]. Everything works as expected, and the ARM build
procedure is now documented in the README.

2. Demonstrated video frame capture on the Raspberry Pi from a
Creative Live! Cam Chat HD webcam via USB [Mike].

3. Demonstrated laser control via a GPIO +3.3V circuit, controlled
from a shell script [Mike] as well as from Python [Arto].

4. Validated the pan/tilt frame's PWM servo motor control via a servo
tester device, establishing the effective range of motion and maximum
slew speed.

5. Demonstrated PWM servo motor control via a GPIO circuit from a
variety of software, including Bash scripting, RPi.GPIO in Python, and
ServoBlaster. Bash proved ineffective (probably due to subtle timing
issues in the inner loop), and Python achieved only jittery control
(likely due to insufficient and/or variable timing precision afforded
by the Python runtime), whereas ServoBlaster demonstrated smooth
control.

6. Attempted to install the Panalyzer logic analyzer on the second
Raspberry Pi board in order to debug produced PWM control signals
[Mike]. Did not work out as yet, due to difficulty configuring the
software for newer Linux kernel versions.

7. Implemented a unit test harness so as to be able to proceed onwards
based on a more effective test-driven development (TDD) modality.

All in all quite a productive first hackathon, despite the limited
time available and the voluminous amounts of inebriating beverages
consumed.

The plan is to organize a second hackathon some 3-4 weeks from now,
most likely again in Bratislava.

PS. Attached a photo of coders hard at work, fueled by local white
wine and BBQ takeout bought with bitcoin. Charlie presiding.

[1] https://github.com/conreality/consensus/wiki/Laser-Turret

--
Arto Bendiken | @bendiken | http://ar.to

[Mike Gogulski]

An impressively productive weekend overall. Even the failures taught
some important lessons, like:

1: Doing pulse-width modulation over GPIOs is extremely problematic
without a C core doing the timing-sensitive stuff.

2: We really need an oscilloscope.

Looking forward to the next session.