Alex headshot

AlBlue’s Blog

Macs, Modularity and More

WinFS is no more -- so what's left of Vista?


According to a post on the WinFS blog, WinFS is no more. Well, actually what was said was:

"These changes do mean that we are not pursuing a separate delivery of WinFS, including the previously planned Beta 2 release. With most of our effort now working towards productizing mature aspects of the WinFS project into SQL and ADO.NET, we do not need to deliver a separate WinFS offering. "

... but in effect, it means that WinFS as it was originally intended (to be a replacement filesystem for the aging NTFS) is never going to be realised. OK, so they've salvaged some of the technology and are going to put it into work in the relational database SQL Server, but since the premise of WinFS was to provide a relational model filing system, what they're actually saying is that they'll be moving relational code into a relational database, which is hardly shattering news.

It's actually pretty disappointing for both Microsoft and also the tech industry in general; other than RAID systems (and possibly journaling), there hasn't been any advances in filing systems for several decades. A mass-adoption of WinFS (which was originally touted as a part of the next-gen OS update Vista) may have moved filing systems into the next century. BeOS showed that BeFS was a great way of storing (and indexing) metadata; every aspect of a file was merely a property of the file in the filing system, which made for easy traversal of metadata such as ID3 tags in MP3 files. To a lesser extent, Apple's HFS (and now HFS+, which essentially is a journaled version of the same) supports multiple metadata streams, but often only the 'resource' fork of an HFS file tends to be used (and in any case, there's no real support for multi-metadata access outside Macs).

Perhaps Son-of-BeFS will hit the OS community, and that will drive the adoption of the next-gen filing system, but I doubt that will happen. Perhaps if Apple wants to make metadata a key part of the filing system (rather than just indexing like Spotlight) then they'll release an open-source filing system that's capable of storing metadata in a bid to make it the de-facto standard. I suspect this won't work unless distributed protocols like NFS are also adapted, which is the real problem.

But the more interesting question is; what's left of Vista? It's been delayed so often, it's starting to become a standing joke in the IT community ("You'll get a promotion ... when Vista ships!"). Not only that, but the pillars of what Vista was going to achieve are falling one by one. Without WinFS, there's not really much to Vista other than a new graphics processing subsystem (Avalon) which unlike OS X's Quartz, only gives a benefit to those apps that have been coded against it. (By comparison, OS X has always used DisplayPDF -- akin to Avalon -- and Quartz and Quartz Extreme have slowly been moving the DisplayPDF rendering to the graphics card over the years, so that all apps are faster by default, without any recompilation necessary. That's one of the reasons why subsequent Mac OS X releases have appeared to get faster, rather than slower, with time -- the processor's load has been reduced by taking advantage of the graphics card.)

So, Vista is essentially going to be just WinFX, as well as a few bundled apps. I don't think people buy a new OS just because of a few apps, and I'm not sure that there's any compelling reason to upgrade. I think that the only reason people will upgrade is when devices start working with just Vista, but even then, there's a lot of old Windows systems out there that hardware developers won't want to turn down (just look at the device drivers you get these days that still support Windows 2000). In fact, the only real devices that are going to be Vista only are likely to be graphics cards, and that's a fairly niche market for gamers who probably upgrade 'just because they can'.

I particularly liked this comment regarding the demise of WinFS:

WinFS is dead. I had viewed Vista as a prerequiste to the eventual release of WinFS. To me this means, there is nothing much of interest on the horizon beyond Vista and there is no real improvement to Outlook coming either. So, if Office 2007 runs on Windows XP, tell me exactly why I want to switch to Vista. Every day the idea of switching to a Mac that can run the few Windows apps I can't do without is looking better and better. Furthermore, desktop Linux probably isn't as dumb an idea as I once thought either. I'm even going to have to go back and check on how Chandler is doing. When a die hard Windows nut likes me begins to see a horizon filled with Macs and Linux desktops with open source core applications instead of the promise of a better Windows, I think that might be something for Microsoft to be concerned about. Ray Ozzie, who I admire, may be Microsoft's last hope.