Script delay

[daveinhull]

Sorry to bring this old topic up again, but I've done a bit of testing.

If I purely use the RFXcom Mgr the button presses are detected and report instantaneously. However, as my previous posts noted, when they are being intercepted by Domoticz, there is anywhere between 1 and 2 (more usually around 2) seconds before they are detected. This makes it pretty useful around the house with the family trying to put up with my 'gadgets', as the misses said, b***** h*** I can't even switch the d**** lights on any more . The problem is the delay causes them to hit the switch again, which then gets trapped and something else (planned for the system, but not for them ) happens.

Anyone with any ideas?

Dave

[jannl]

Learn her to be a bit more patience
I do not notice a lot of delay when operating a switch. An automatic action triggered by a x10 pir can take a second or so sometimes. Guess I to get me a pi 2...

[theogough]

frankly the poor latency is my most significant issue with domoticz, which otherwise is pretty decent and has good community support.

i've got 'this close' to full replacement with something that can genuinely be reactive to rfxcomm's event stream. i havent had a chance recently to review the code but my suspicion is that all of this is down to a polling read of the tty stream from the rfx receiver, as opposed to a faster approach of reading an open stream and pushing events into domoticz as soon as they are received.

i admire various posters attempts at finding relationships between certain lua script forms, or combinations of hw protocols but i fear the root cause is something far simpler.

i suppose next step is to review the code....

[alfred_j_kwak]

I have switches controlling lights and use Domoticz sometimes to control switches.

[theogough]

@daveinhull i feel your pain! the latency is the ONLY reason i get it in the neck and have to dial back the amount of automation the house could be doing. eg i wanted to have the downstairs shower/toilet light activated by pir - NAH too slow. switched on manually (using an LWRF stick-on master switch) then timer to turn off if not done manually - NAH too slow.

i have almost literally had to avoid getting something in the neck each time i've activated something that has what amounts to unacceptable delays.

guys,, be honest - 1-2 seconds is JUST UNACCEPTABLE! it is NOT FEATURE EQUIVALENT with a hard-switch and you shouldnt have to TEACH people to wait SECONDS!

The fix is to target LATENCY as the primary non-functional concern of the entire system. anything more than about 400ms is going to be annoying at the human level, and get real - this system should have HUMAN interoperability as its central tenet.