What is the recommended migration path from 2.x to 3.0 with KNX devices?

Dear Community,

I’ve updated my docker based openHAB system from 2.5.10 to 3.0.0. All my things, items, groups and my default sitemap in basicui are usable. Switching from Joda Time to the Java Time API was easy, so also my rules are working.

So it seems to me that going for semantic model would be the next migration path to go. Currently my openHAB system mainly is serving a webview on my devices and adds some higher logic functions. Most other functions are either done via KNX or HomeMatic.

Regarding the semantic model. What is the recommended way of assigning locations? Would you go for a logical or a physical view? This question comes up to me in the context of the following examples:

  1. My rain sensor is a dumb one with only a switch contact. The connection to my KNX system is via a 16 channel binary input module. So the physical location of the sensor is my electric control cabinet whereas the logical location is the rooftop. What is the preferred way to put this in the semantic model?
  2. Nearly all of my actors are also located in the electric control cabinet. Same question here. What would be the preferred assigned to locations?

In my openHAB 2.5 installation I kinda used both views.

The physical view was represented in the paper ui. Nearly all actors where located in the electric control cabinet and therefore also the channels, because the view was derived from the physical model of the KNX bridge. And the naming convention followed the electrical circuit plan done by the electricians.

The logical/functional view was represented in the sitemap with items and item groups. All rules with higher logic functions worked on items or item groups.

Thanks in advance

Read the New User Tutorial and watch the meetup presentation video.

On your title question: there is no such thing as “the” recommended implementation advice on whatever. Anyone’s implementation is unique that’s why.

1 Like