Windows Longhorn: No Bull.
Longhorn, aka the next version of Windows, won't ship until late 2006 (if then), but it has ramifications that begin today. Recently we spent a day at Microsoft learning more about this new OS, and I'm convinced that this is going to be a tsunami-size wave. Smart companies need to fit Longhorn into their product roadmap ASAP. Here's why:
A little history.
The evolution of the PC and its operating systems has been a classic example of stepwise refinement. DOS was basically just a port of CPM to the Intel X86 architecture. Windows was just a shell that sat on top of DOS, and Windows 95 was just a better shell, still sitting on top of a bunch of old 16-bit DOS code. NT was new, but it couldn't handle most of the legacy apps that Win95/ME/98 could. XP saw a merger of the two code bases into one, so the only difference between Home and Professional was a feature list.
Along the way all the 16-bit code went into the bit bucket, the UI got revised several times, more functionality was added, and we saw the object model evolve from DDE to OLE to COM, COM+, and DCOM.
Along comes Longhorn.
Longhorn takes some big leaps, primarily in three areas:
- The file system (WinFS)
- The file system gets completely redone. No longer based on
NTFS or
FAT32, in effect the whole file system becomes a
database. Instead of files being monolithic data stores (think Outlook pst files), WinFS stores items,
which are defined by the application. Items bring two interesting changes: granularity and metadata.
Granularity means that very specific pieces of application data can be stored separately, instead of munged into monolithic files. Each item has associated metadata, such as author, subject, date created and revised, location, and any specific metadata the application wants to keep track of. For example, a digital image might store the subject and EXIF data as metadata, allowing you to search from within the file system for images with certain subject attributes. Now searching for items with subject “Jones Project” might bring up Word documents, Powerpoint proposals, individual email messages, a Siebel or ACT account record, customer invoices and payment records, etc.
- The presentation system (Avalon)
- You heard it here first: this is the coolest thing you've seen yet on a PC. The graphical view you have right now is bit-mapped; with Avalon it moves to vector. Instead of just throwing bits up on a monitor, Longhorn will use the graphics processing unit (GPU) to draw vector paths. Now fonts (and pretty much everything else) can easily zoom in and out. Imagine that zooming on an icon in the file system actually shows you the contents of the file. High-performance animation—similar to today's Flash or Shockwave movies—will be simple to develop and integral to the way information is presented.
- Transaction services (Indigo)
- COM is dead, Web Services are in. Sort of. This stuff is pretty geeky,
but the short, oversimplified version is that Microsoft recognizes that .NET has some severe limitations
as an object model. With Longhorn, Microsoft moves from a distributed object model to a services model,
with inherent support built right into the OS. In fact, WinFX,
which replaces the Win32 API as the programming model, is totally based on .NET.
The main takeaway for Indigo is that creating—and consuming—secure web services will become dead simple. This opens the door for component suppliers to start selling lots of componentized software that's easy to access and implement—something that's been woefully lacking up to now.
What the heck does all this mean, anyway?
Think back to the first versions of Windows. Lots of independent software vendors (ISV's) just put a Windows shell on top of their old DOS apps. Smarter ISV's rewrote their DOS apps to take advantage of the graphical user interface (GUI), multi-tasking, and interprocess communications, limited as they were. Those apps out shined the simple DOS-port approach, and many of those vendors went on to great successes.
Longhorn will be no different. Your existing XP app should run fine under Longhorn, but it won't be able to take advantage of most of the new cool stuff unless you do some significant modifications. Microsoft, as always, will want to showcase those ISV's that have great Longhorn support when the OS is released, giving the ISV's plenty of high-value exposure. It's time to look at your product roadmap to see what you want to do in 2006. And remember: fortune favors the bold.
