[OpenTRV-dev] RF protocol

[Stuart Poulton]

On 20/03/13 10:57, Jack Kelly wrote:
> I'm a computer science PhD student at Imperial (unfortunately this doesn't
> mean I'm a coding ninja!  Far from it!)  I'm co-supervising two MSc
> individual projects on "smart heating controls".  These projects will run
> from early June until early Sept.  Both students are excellent.  One of the
> students (Shubhangi) is interested in using machine learning to predict
> room and house occupancy (to automate heating schedules).  The other
> student, Sokratis, is interested in lower-level stuff and has mentioned
> that he is really interested in building a general purpose home automation
> / sensor protocol.
>
> Which is a very long-winded way of saying that, if you want to, you should
> discuss your ideas for a home automation / sensor protocol with Sokratis
> because this stuff might be right up his street.  I'm seeing Sokratis later
> today so I'll discuss with him and get back to you (unless you're reading
> this, Sokratis?! In which case feel free to jump in, of course!).
>
Ok. Having been involved in the design of a home automation protocol by 
committee, I'm going to impart the following.

1 - It's hard
2 - The more people that get involved the harder it gets
3 - Even when you think you've nailed it someone says what about......
4 - People decide to fork the project.

and now my current view.

We've got some good ideas, based around hardware and software.
Lets get some markers in the sand using existing hardware, and 
simplistic protocol, then look at improving the software side, at the 
moment its still in the techy, nerdy, hacker space, we can all flash new 
firmware.
Once we're able to show things working and in use, nothing worse than 
becoming a 'kickstarter' type project that talks but doesnt' deliver.
KISS is the key, "Keep it simple & small"

Stuart