Sorry my first post is a collection of disparate questions, although I'm very inspired that MySensors.org is the exact place to come to tie all of my ideas together so here goes:

I'm about to renovate my home. We are gutting the interior - blank slate, CAT6 can go anywhere and everywhere.

So I'm trying to decide whether to hook up my nodes using the NRF24L01+ OR whether to go CAT6 on the basis that ... I can. And I think it will always be more secure / reliable. (Even though I don't doubt the security and reliability of RF...)

I've speculatively purchased ethernet shields, nanos, a tonne of NRF24L01+ and a whole load of sensors.

1/ First question. What are the considerations when choosing whether to go RF between a node and a gateway, or CAT6 between a node and an automation server? Will MySensor code available here work with both? Is it that generic / adaptable? Can the nodes themselves speak MQTT directly if I hook them up directly to the network? Or does this require a gateway?

Or perhaps I should still go with the node-gateway architecture using MySensor code to make use of the adaptive nature of node detection, but still use CAT6 to connect nodes to the gateway? Is that a 'supported' scenario? (Sorry, the word "supported" is silly, but I just mean has anyone else done it!)

2/ I'm in love with OpenHAB. Absolutely love it. It ties together so many other areas for me, because of the sheer amount of bindings that have been written.

If I use RF between nodes and the Arduino gateway (as suggested on this site), can this gateway communicate over the network (using a network shield) with my OpenHAB server using MQTT? Does code exist to do this? I don't mind hacking existing code but I'm a bit nervous jumping straight in and trying to code MQTT to OpenHAB myself!

3/ I guess most people here want to retro-fit their wireless sensors, therefore are happy working with battery powered Arduinos which sit directly next to sensors like the DHT22.

I want to flush-mount ALL of my sensors, and never change a battery in a node. But I still want my nodes in an accessible place in case they malfunction. This means I want permanent installation of sensors into ceilings, but accessible nodes (e.g. wired into a cupboard in every room).

What's the maximum length of cable I should use to connect the sensors to the nodes? I'm thinking CAT6 cable for e.g. 3 sensors in a ceiling back to a nano. Then another CAT6 to another sensor in a different place back to the same nano. Approximate distances of a given run would be 5-10m.

I know it depends on the sensor, but does anyone have experience of situating the nodes away from the sensors like this?

4/ I want to make a room audio control unit, using an Arduino subscriber of MQTT from to a publisher with binding on the OpenHAB server which is connected to Squeezebox. It will have a read-out of the current song on a display, and have a few buttons to trigger play, pause, next, and a rotary encoded for volume.

Anyone done anything like that? Does there exist any code to do such things, perhaps again with a little hacking to get it working for my scenario?

Sorry about the stupidly fragmented thread but hope to hear some answers here :)