Selective Recalculate

[skyewright]

There has been some talk recently about increasing efficiency, how about this...

I have just swopped over the channels used by 2 of my OS sensors.

Meteohub thinks it needs to do a recalculate (though in reality nothing has changed - the same sensors have the same MH IDs and names, they just use different OS channels).

Even if a recalculate was relevant, it would only be those two sensors that might be affected. Nothing else has changed at all.

Similarly, if I added a calibration factor to one sensor MH would currently want to recalculate everything.

So, how about if there was either extra intelligence in MH to only recalculate the relevant sensors or perhaps some setting that allowed the user to decide which sensors were to be recalculated.

Presumably just recalculating certain sensors would be a lot quicker (and less stress on the NSLU2) than recalculating all?

A possible interface for setting which sensors to include in the next recalculate might be a series of check boxes on the Sensors page?
Maybe all the check boxes would revert to 'yes' after a completed recompute (so the default was everything), maybe not?

What do you think?

And since I'm in wish mode, perhaps when a change is done that would currently force an automatic recalculate it would be very nice if instead it just warned me that a recalculate was required and that accumulated data values might be invalid till a recompute was complete (maybe setting System Info - Weather Data Recomputation to say "Required" too) but leave it to me to decide when a recompute would be convenient.

[skyewright]

I am even more keen than ever on this now.

I made the two mentioned changes one at a time, but within a couple of minutes of each other (i.e. reset one sensor, then moved the ID and name to the new signal, then repeated that with the other).

MH has decided that requires two whole recalculations!

As soon as it had finished the first full recalculate of all sensors it started a second recalculate!

Each recalculate takes over 4 and a half hours during which time I need to reduce general activity. So that's 9 hours of high processor load - at the end of which the accumulated figures will be just as they were before it started.

There must be a better way...

[Chill]

...Each recalculate takes over 4 and a half hours during which time I need to reduce general activity. So that's 9 hours of high processor load - at the end of which the accumulated figures will be just as they were before it started.

There must be a better way...

I agree! Actually changing interface language also activates a recalculation... No sense! Neeeds to be improved... Isn't it?

Chill