[OpenTRV-interest] bug report: binary format light scaled incorrectly

[Damon Hart-Davis]

Hi,

Thanks: filed in JIRA as TODO-510.

Note that I have done a fair amount of dithering with the scale of the light value for various reasons (such as transmission bandwidth, a tiny bit of privacy, and a partly-hidden entropy source).

I’m not astonished that I’ve broken in.

Note that representation on the status line may or may not match elsewhere, eg different transmission formats.

Do you know if the problem you see is unfixed in the HEAD version of the code?

Rgds

Damon

> On 22 Mar 2015, at 08:47, EMAIL ADDRESS HIDDEN wrote:
> 
> Here is a sample taken from OpenTRV output
> 
> @2532;T20C4;L35;O1
> FHT8V TX
> =W0%@20C3;X0;T21 5 W255 0 F255 0 W255 0 F255 0;S18 7 18;C5;HC37 50;{"@":"2532","L":141,"B|mV":3179,"occ|%":0,"vC|%":0}
> 
> Binary format light is scaled incorrectly i.e L35 versus "L":141.
> 
> This problem seems to have arisen since the basis for light reporting changed from [0,1023] -> ~[0,255].
> 
> ASIDE: In a side project and simulation of the opentrv algorithms and variants to identify the use of electric light I have noticed that the light lsb's are borderline important to identifying the use of electric light. Whatever the internal opentrv algorithms, reporting of all measured bits is most likely important to measurement, analysis, and subsequent simulation of algorithms and/or variants.
>