How long(horn) blues

There is a wonderful Scots band called The Blue Nile, whose music I highly recommend. However, it takes them years to get a new album ready. Whenever one actually arrives, it is almost a shock. Well, it is getting like that with Longhorn. This was unveiled some 18 months ago by Microsoft at the Professional Developers Conference, and there was much to like – some of it undoubtedly brilliant – so we looked forward to the next instalment. Only the most dedicated of Microsoft-oriented developers kept Longhorn on any real hardware, though, so it was with some surprise that it turned up again at the recent WinHec conference in the US. I would love to be able to tell you that it is finally up to beta quality, but it is far from that. However, it is clear that the engineering teams have been working hard behind the scenes, if only to sort out the mess of premature announcements, pious wishes and pure hope that was expressed nearly two years ago.

As you would expect, Microsoft is trying to push hard on the desktop graphics front, but in doing so it knows it cannot afford to leave its customer base behind. While it would be nice if everyone had super-high-end graphics cards, the stark reality is that many graphics hardware implementations have been moving toward shared-memory solutions that offer little hope of performing well in the realm of high-speed 3D graphics – this is especially true in the laptop arena. At least new technologies like PCI Express appear to be gaining a significant foothold in the power desktop market, pushing aside the now inadequate AGP solutions

While most of us will probably wait for Longhorn to ship and then buy new Longhorn-compliant hardware, the Microsoft suits want it to run on anything and everything that’s out there now, which obviously poses an insoluble problem for the graphics team. The solution that they have come up with is to split the graphics experience into four levels. At the most basic, there is the Classic environment, which is similar to the existing XP look-and-feel and is clearly aimed at lowest common denominator hardware. It is intended that Classic will support two display-driver models called the XPDDM (XP Display Driver Model) and the new LDDM (Longhorn Display Driver Model), again emphasising that this will be the plain vanilla version of the OS to be used for corporate upgrades and lesser hardware setups.

Next up the scale will be the To Go version, which will look and feel like the next higher Aero Express version, but will not have any of the latter’s advanced features, such as multilayer video compositing. I expect this to be the minimum level Microsoft would really like you to run at, to at least get something that visually looks like Longhorn. Then comes the more mainstream Aero Express version, which is what we have been seeing in the demonstrations so far. This will require Longhorn native graphics card drivers and thus a minimum level of graphics processor capability. It should do some scaling, animation and so forth, but it will not be the very top – that role is reserved for the fourth level called Aero Glass. Little has been shown of this version as yet, except that we know it will require a top-end graphics processor and lots of RAM. It will display the premier version of the UI, allowing for things like transparency in multiple-layer compositions, including multiple layers of video composition. The word is that we will not see much of this product until a mid-beta release, possibly next year. If you are a display driver writer, you need to get your skates on and start working on Longhorn Display Driver Model drivers right now, but in truth most of the drivers will come from the mainstream hardware chipset houses that write generic drivers.