Not only the number of devices, but also the complexity of the integration makes the problem of remote control complicated. I gave up trying to find the optimal programmable universal remote control of the conventional type. Having spent hours on "programming" the Lapeschi Telegenius, and then losing everything with a battery change or a firmware failure — enough!
Therefore, I bought a Philips Pronto RU 890 (now obsolete since a long time). This magnificent gadget comes with a touch screen, which can be (almost) arbitrarily filled with customized device- and macro-pages. Programming preferably takes place with the ProntoEdit Windows program (download-able from Philips free of charge), or from the free software Tonto. The program is written entirely in platform independent Java, with sources available (Unfortunately, the program does not support the newer Pronto models, and its author, Stewart Allen, discontinued its development in January 2004.) There is a large user community for the Pronto (and its "compatibles"), in particular, as a great source of inspiration, knowledge, and files, remotecentral.com should be mentioned.
With the Pronto, a completely field of possibilities was opened. But it does not offer a quick solution to a problem, rather presents you with a new hobby :-).
Using the RF transmitter and the Pronto, I analyzed the signals, and was able to come up with a C++-program that generates the Pronto codes for arbitrary signals for the Conrad stuff. Here are the (GPL-licensed) sources. To compile, use a command like gcc -Wall -o conrad conrad.cc -lstdc++.
In my living room, all the low-voltage halogen setup, as well as the TV background light, are all switched by the Conrad switches. Also the fountain, water column (powered by an aquarium air pump and halogen light), and water wall are switched by these switches, and thus from the Pronto.
Some A/V devices do not come with remotely controlled power switches that do what I want. This applies to the laserdisk player, the sub-woofer, and the MK2 CD-player. These can also be switched on and off through the radio controlled power outlet strips.
Infrared codes, in particular discrete codes
For the IR hacker, searching for new commands is like searching for comets and supernovas for the astronomers. In particular, many functions on today's consumer electronic is only available through so-called toggle-codes: For example, pressing the Mute-button turns the muting on (if it was off before), or off (if it was on before). When a human is pressing the buttons on a remote control, this is of course not a problem, because the human knows the state the device is in. If he doesn't know it, he is probably not interested in changing it anyhow. Things are different if a "machine" (for example a macro on a remote control) instead of human is pressing the buttons. Pressing "Power (toggle)" may turn devices on, instead of off, or vice versa. Commands that put a device in a defined state, like "Power on" or "Power off" are called "discrete codes". It should be obvious, that devices with no discrete commands, just toggle commands, lead themselves very badly to automatization.
A look at todays consumer electronics shows that most devices are in fact TOADs ("Toggle Only Actuated Device"), instead of DUDE ("Device Utilizing Discrete Extensions"). In many cases, undocumented discrete codes do exist, and discussed in forums like remotecentral. (Developers are often smarter than the people selling their work.) In some cases, they seem not to exist. Why is this? Of course, the number of physical switches on a device increases the final cost, but this is not the issue. Buttons on a remote control cost less, but has a frightening effect. Still, it is hard to find a good reason for not offering any, e.g., discrete power commands. Particularly disappointing are the magazine tests, for example in German magazines. "Discrete commands" is obviously not a part of their tests, not even their terminology...
With this device, in late 2003, Panasonic manage to produce an absolute price-performance star. It is so much more disappointing, that with respect to the issues here, they managed to do almost every possibly error conceivable. No discrete power signals exist. To select a particular aspect ratio, it is necessary to click yourself through a cyclic (!!) sequence of possible aspect ratios. (It it was not cyclic, it would be possible go to the end of it ("up" many times), and then have a well-defined state!) Discrete input selection is also not possible, but have to be selected through cyclic sequences.
I analyzed the signals, and wrote a program to search through 256 possible candidate codes. Nothing interesting was found. Also on places like remotecentral, nothing appears to be known. The name "Panasonic" does not appear to command a great deal of respect there...
For this reason, I doubt that my next projector will be a Panasonic...
For Yamaha, similar things can be said, just with the different sign... Many, but not all, commands are discrete on the standard remote control. But more commands exist, even documented (!!) on this page. Yamaha even provides its users with a tutorial on how to use their extended codes!
Ongoing work; future setup
The possibilities of the 1MB Pronto are since long exhausted, and I am working on a new solution. Exactly how this will look like is still unclear, however, componentes will probably be a Siemens Simpad SL4, (probably running Windows CE 4.2 from Simon Mullenger), using Netremote 2 from Promixis, together with Eventghost running on the HTPC. I acknowledge the influence of this thread by Beisammen. Watch this space in the future...