Or one could consider that only very recently technology has become cheap enough to be a viable option. In fact, I'm still not convinced it's cheap enough to interest the average consumer. Tho Apple's team probably disagree.

What do you mean by not plug-and-play? Any BACnet device can be connected to a network and announce itself with a WhoIs, being instantly recognized by all the other devices.

(...) supports a relatively small number of transport protocols (...) BACnet has many official transport protocols, the 3 most popular being MS/TP, Ethernet and IP. Furthermore, nothing prevents the BACnet packets from being embedded into some other protocol.

And because you mentioned ZigBee: http://www.zigbee.org/Standards/ZigBeeBuildingAutomation/Ove... "The ZigBee Alliance joined forces with BACnet, the leading global building automation and networking protocol for building automation, to fully support BACnet over ZigBee Building Automation networks. Soon it will be possible to easily expand your wired BACnet-based building systems to new areas by using wireless ZigBee Building Automation products."

You are right about security, BACnet was not designed with security in mind (nor was the Internet). But this brings other questions:

1. Does the consumer space need security for thermostats?

2. If it does, why shouldn't this be handled by the network security?

3. If there's really a need for separated security, why not add the required changes in the next protocol version?

(i.e. most BACnet/IP devices will stop working as networks switch to IPv6)). No. No they won't. Firstly because most BACnet network are to remain inside the enterprise own intranet which will remain IPv4 for a long time. (Forever?) Secondly because IPv4 is a subset of IPv6.

BACnet is probably fine in the enterprise domain, where everything is installed at building construction time(...) Which is simply not true. Most buildings are many decades old. They are a (horrible?) mess of pneumatic systems, first generation electronics control with relays panel, and what we could call recent technology controls with micro-controllers.

In the home domain, it's at least overkill, and makes things incredibly difficult for users. I fail to see how using existing technology is 'overkill'. Was it overkill to bring Ethernet and IP to home users? Should we have come up with a different standard?

(...) Trying to do a standards-based approach at the moment would be a horrendous idea - firstly I know a large number of companies to be trying to come up with their own solution at the moment (...) How is many companies trying to come up with their own solution an argument against standard-based approach?

(...) but secondly it would just end up being a standard by committee (with no real implementers). You mean without the worldwide HVAC industry, right?

Personally, I find the BACnet protocol to be bloated. But is it a good enough reason to throw away everything?