I’ve just returned from Intel’s Developer Forum in San Francisco. I went expecting some big noises over next generation IP-based television-cum-other media-cum-internet enabled systems as well as some in-depth and staggering information on Intel’s new cloud approach.

OK, there were various Smart TV demonstrations, with the vast majority just looking like marginally advanced set-top boxes, and many this generation TVs with embedded internet capabilities have close to the same functionality.

Wireless Display (WiDi) technology looks promising—this makes it possible for the guts of a next generation TV to be placed into a box and for one or more displays to be driven by the box. Now, for a full centralised system, this does bring up some interesting options—imagine a TV where an incoming video call can be directed to the main 42” TV screen while the programme is paused and recorded, or to a smaller personal screen so that the rest of the family can carrying on watching the latest blockbuster 3D film being streamed to them over metro Ethernet.

Imagine being able to pause a film that you were watching that you thought was only an hour and half long only to find that it is two and half hours, move from the living room to the bedroom and pick up where you left off. Maybe being able to carry out a videoconference that tracks you from room to room if you so want—all possible with WiDi, if the Smart TV systems can be created to deal with the detail—and this will need centralised systems that play into Intel’s hands.

There were also some discussions around cloud—but little that would grab businesses by the short and curlies and make them want to go to Intel for more information and help. Intel has some issues that it has to deal with. Firstly, it wants to be completely bought in to heterogeneity yet this would mean that it can’t do anything at the silicon level that isn’t supported by AMD. It can remain agnostic above the hypervisor, though, but where would Intel then add sufficient business value?

If it tries to move too heavily into the systems management market, it will not be seen as a major player, and will find it difficult to face down the likes of IBM, CA, HP or BMC. If it provides tools that make the most of Intel’s own silicon, then it will need to make sure that the tools fit well with existing systems management systems and that the value to the business is clearly messaged.

Intel’s deeper approach, where it is looking at how it can play as a project manager for datacentre builds, is a far more interesting area—yet it wasn’t looked at in any real depth during the sessions I saw. A fully balanced view of how silicon fits with networks, facilities and software is needed when it comes to next generation datacentres—and I do believe that Intel can do something solid here, if it can put together the right teams.

So, what did grab me? Firstly, Intel’s Atom processor is to be packaged with a field programmable gate array (FPGA) in a single package. Why is this important? In its first instance, it is probably less of a major step than I expect it will be in the future. Currently these two discrete items are just being put onto a single pluggable system. There will be some speed benefits, and code to reprogram the logic of the FPGA and the CPU firmware/operating system can be done in one place.

However, if/when Intel goes the whole hog and goes for complete integration, the use of cpu/FPGA composites could change how appliances and devices are built. As codecs, security algorithms, network standards or anything changes, appliances and devices could be upgraded in situ directly through software without the need for costly and impactful hardware forklift upgrades.

Lastly—and my favourite—was something I happened across. A UK-based company, RealVNC, provides software that allows computer screens to be replicated on devices anywhere over the internet; a sort of infinite keyboard, video and mouse (KVM) solution if you please. I’ve been using VNC for years, and it has saved me many times when applications have failed on my desktop or server—just log in and take control and off we go again. However, if the machine itself had collapsed and hadn’t rebooted, then I was stumped.

RealVNC has worked with Intel to drop the capabilities directly into vPro. Now, provided that the machine has power to it and there is an internet connection, you can get through to the machine. No operating system is needed—you can get through to the BIOS and configure that, you can even install a different OS if you want while away from the machine. Help desks will be able to take control of a user’s machine and carry out diagnostics right down to the hardware level even if the OS has corrupted. To me, this has a load of business value—and yet Intel relegated the announcement to a 3 minute slot in a rushed “Newsapalooza” session with only the press present.

One small caveat—I was just one person, and IDF was a massive event. With the best will in the world, I could only cover a small amount of what was on offer. However, the main tent sessions, where the majority of people were, are surely the places where Intel should be setting out its stall in the best possible manner to attract not just the developers, but the analysts, press, end users and those responsible for buying Intel-based solutions.

Intel has a lot to offer—yet still seems to fall back on to technical messaging at the earliest opportunity. Genevieve Bell, Intel Fellow and Labs Director Interaction and Experience Research, put it best—Intel has to change from looking at a TV and wondering how to get a computer inside of it. Based on IDF 2010, I think she has her work cut out to persuade the rest of Intel that she has hit the nail fairly and squarely on the head.