If I may weigh in. A few main points on MDs and why as a user I continue to use them and would be disappointed to see them relegated to a unsupported system.
Number one for me is Cost. An ip controllable HDMI matrix, HDMI endpoints, multiple set top boxes/STBs = multiplying the cost to every-room and every zone you add. You can add 'smarts' to any display with a relatively low cost MD and a remote.
I enjoy the ability to control any zone/room from and zone/room without having to have a dedicate room controller (ie. iPad) in each room. Call us 'old skool' but we also prefer to flick through with a remote rather than always use an ipad, ipod touch etc.
The philosophy of LinuxMCE/Dianemo is the central home control. Not just single point to point but in addition multipoint and the ability to move that media/user experience between rooms seamlessly. This includes the same UI, the same control process and access to all media in all rooms. As well as lighting, heating etc. controls. In your model, the only UI is on the mobile orbiter.
Telecom, Game player and surveillance cameras are not considered here yet within the thread. Whilst I agree that the most used component of the system is media, it is the integration of all these smart home systems in one point and turning any display device into a controller and viewer that got me so enthused in the system.
In having said this, not all times is there a need to have a UI and consistent user experience. I use media players and upnp in conjunction with squeezeslaves and traditional MD to get a mix throughout the home. I would just hate to see them become unsupported.
Food for thought.
All good points but I think you have misunderstood some of my comments here and in other threads. So let me try and clarify things;
In Dianemo & Linuxmce we have an incredibly flexible tool. We can organise and structure our installations using many different approaches/strategies. Some of those can include centralised resources like STB's with HD video distributed to each room via IP Controlled HDMI Matrices. Some might prefer no centralised traditional STB's but instead use Myth or VDR backends running in the Core. Some might prefer STB's in each room - but that is very rare indeed. But all of these approaches and many, many more are possible.
With today's TV's you already have 'smarts' in the TV - just plug a network connection into the TV and it can be controlled from anywhere by any Orbiter. If your TV is slightly too old to have an on-board Network connection then install a Rpi running RaspBMC behind it and you can get all the same capabilities, including CEC control, on your existing TV.
Orbiters provide the ultimate in flexible control. Yes you can rely on on-screen Orbiters for all control but that means you need a TV in each room to access the UI - and what about those rooms where you may not have a TV? With portable devices we can take our Orbiters to any part of the house and control that room, and we can still control any other room while we are there. I addition we can do that without having to have MD's and TV powered up to do so.
With Dianemo or LinuxMCE we have centralised home control - but is does not require multiple MD's + TV's all over the house to achieve it. Every device type that is supported by either system can be integrated, controlled and accessed without a single MD - they are not required to achieve that at all. But you can have an MD in any room if you want it with an on-screen Orbiter - mix & match without limitation. Orbiters running on portable devices or on PC's can be used without limitation to control any room and any device anywhere in the system. You don't need, and I have never seen, an installation where you need a portable Orbiter dedicated to each room. You take your Orbiter with you or use one that is already in the room.
There are two scenarios where that require an MD;
- Live TV streamed from Myth or VDR in the Core.
- Using the Game emulator
Everything else can be achieved without MD's and without any loss of capability, flexibility or centralised control, while at the same time reducing complexity and energy costs. But in the end of course it's your decision - that's the ultimate in flexibility.
All the best