Slashdot Mirror
WinFS Gets the Axe
commander salamander writes "Over at the WinFS Team Blog, Quentin Clark states that Microsoft no longer plans to ship WinFS as a standalone software component. Instead, portions of the underlying technology will be included with the next release of SQL Server (codename Katmai) and ADO.NET. Does this spell the end for the true relational storage paradigm that Microsoft has been promising since Windows 95?"
Yes.
If they didn't put back WinFS, they couldn't use it as vapo^W a feature of their next product. And when that product comes out, they'll push it back to the product after that, just like they've been doing for the past seven or eight years or so.
WinFS is the perpetual motion machine of vapourware. They are constantly promising it for their next product, but they never seem to deliver. That doesn't stop $NEXT_PRODUCT from being compared favourably with the competition because of WinFS by PHBs though.
Bogtha Bogtha Bogtha
Their structured, indexed filesystem that operates much like a database, will be released with their database software!
Is it just me, or does that sound slightly redundant?
It's better to vote for what you want and not get it than to vote for what you don't want and get it.
- E. Debs
Of course, what do you think the whole Vista release is? It's windows XP + more CPU and RAM usage. Nothing special or useful. It's pretty pitiful that developers being paid 6 figures work at a slower pace than volunteer open-source developers. It also fits into Microsoft's motto: Less Later. Just like Halo 2, one of the most anticipated games was out years late and really sucked.
Microsoft will always do this, just like Vienna (Fiji, whatever) is supposed to be a complete re-write, bullshit. They'll probably just add some crappy RAM and CPU hogging features and call it inovative.
a. WinFS had difficulty functioning over a network
b. Microsoft's target customer is business
c. Businesses use networks
Therefore, WinFS would not be suited for business usage, making it unimportant.
Hey, if everyone wants to bag on Microsoft not making a next generation file system, what is stopping Linux and the Open Source community from doing it? Oh, that's right- it's easier to just complain about MS than to actually get your hands dirty. Nevermind then, carry on.
"Does this spell the end for the true relational storage paradigm that Microsoft has been promising since Windows 95?""
Were's the F/OSS equivalent?
Simply call it what it is, fraud, from a company that knowingly engages in deceptive business practices.
To be fair, I would call this a failure rather than a fraud. They probably believed their own claims about WinFS at the time that they made their promises.
-jcr
The only title of honor that a tyrant can grant is "Enemy of the State."
And meanwhile ReiserFS on Linux provides much of the functionality today that WinFS only promised for the future.
Shh.
On the contrary, plenty of corporations might lie, but how many companies can get away with telling the same lie over and over and over again?
"Yeah, sure, WinFS will be in this one. It's not like last time, or the time before that, or the time before that. We mean it this time."
Bogtha Bogtha Bogtha
From what I remember the funky file system was the last piece of the Cario product/suite of technologies to need to ship since Allchin announced it in 1991. Sadly for Allchin, they were unable to deliver on the last great promise of Cario before he leaves, and the poor guy waited 15 years for it too.
$2B OR NOT $2B = $FF
What's really sad is that BeOS had a woking usable dbFS TEN YEARS AGO!!!! I bet Visa idles more RAM and CPU resources than an BeBox had to begin with.
Hell, it makes you wonder if they had an idea they weren't going to ship even as they demoed WinFS at TechEd just two weeks ago.
And just think, enterprises rely on this company's OS, which is so internally complicated that its own developers call it "broken." It's amazing the economy came to rely on a company so unreliable.
"Sufferin' succotash."
Why do microsoft bods keep using the term 'rich' to descibe their technology?
Most notably, how is it that they seem to apply it to technology that never gets to the production stage.
It's almost as if they feel it aboo to admit that their technology is untested, nay imaginary.
I don't care if they have some in house code. If it isn't in circulation, it's not technology, it's a unproven concept, and definatelly not 'rich'
It's the opposite extreme of tight-lipped Apple, who get criticized for not providing a clear roadmap*.
Whether this dropping of a major technology is due to inability to implement or typical MS marketing strategy (don't buy our competitor's product, wait for ours!) doesn't matter at this point; MS has been overtaken and it's only a matter of time before the world completely passes them by.
At one point, I thought Linux-on-the-desktop had a limited window of opportunity to reach the point where ordinary people would embrace it before "Longhorn" came out. I think MS has effectively welded that window open.
It's not offtopic, dumbass. It's orthogonal.
How many articles have we read that tell us that the boys over at Redmond lack organization? There's the famous story about the two Office development teams that built two versions of Office with incompatible file formats, because neither team knew about the other. There's all the stories about managers being forced to lie to make time table deadlines.
Now, all of that boils down to one simple thing: The left hand REALLY has no idea what the right hand is doing. What makes you think that their marketing team is any different?
Its easy to point the finger and cry that they lied, but is it really a lie if they didn't know any better?
"Copland is to Mac OS 8 as Longhorn is to Vista" seems to be becoming more true every day.
Though it was promised as a fundamentally ground up re-invention (Pink, Copland, System 8), the Mac OS 8 product that was actually shipped was mostly a cosmetic upgrade with the bits of the promised technologies that could be made to work. The new graphics architecture became a new font subsystem. The new document archicture (without developed parts making use of it) became a built-in web architecture. System wide document content searching became better file finding. The goal became to try to keep whatever anticipation was already built but jettison the "hard problems" of making it actually work in the ways that were promised. Tell everyone that Feature X has evovled into something beyond what we had ever anticipated rather than the world passed us by while we were shooting for an old target.
It may be that Microsoft still has the inertia to pull off an almost completely cosmetic update, but it's going to get pretty ardurous environment on the development teams. After all, the goal isn't going to be to even ship a feature reduced product. It's going to be to ship cosmetic filler that covers up the need for what was really promised. Maybe Blackcomb or Fiji or whatever it's called now, will become a stage for the proper solution, but that's a very big IF.
I hope Vista will come with some serious eye-candy
If the betas are any indication, you'll be disappointed.
-jcr
The only title of honor that a tyrant can grant is "Enemy of the State."
Fraud? So you lost a lot of money due to this whole WinFS thing? Did it kill your children or something?
Linux already has different filesystems that have implemented the "features" of WinFS for decades. Take ReiserFS, JFS, EXT3... they are all journaled database-like systems. Even some engines in MySQL can do what WinFS wants to do. What Windows REALLY needs is native support for said filesystems, so they can go on with the rest of the world.
Custom electronics and digital signage for your business: www.evcircuits.com
So ext3 was a clean sheet re-do? Isn't it backward compatible with ext2? Indeed, isn't it just ext2 with journaling?
Always thought this would happen. Not only was the original concept of WinFS pretty difficult from a technology point of view, but people at Microsoft suddenly thought: "Hang on. If we deliver a rich database storage engine integrated into Windows then that threatens the existence of SQL Server." This is confirmed quite adquately by this:
"We are choosing now to take the unstructured data support and auto-admin work and deliver it in the next release of MS SQL Server, codenamed Katmai. This really is a big deal - productizing these innovations into the mainline data products makes a big contribution toward the Data Platform Vision we have been talking about."
Notice the word 'productising' (productizing for you yanks). Productising here means "Why give this away for free in Windows where it would actually threaten the existance of SQL Server when we can just bundle it into the next release of SQL Server and charge people more for the *new* features?!". This is confirmation, if ever it were needed, that WinFS is totally dead as a Windows component. You're not going to be able to tag your files, or 'objects, with metadata and search for it seamlessly along with new integrated and built-in Windows file management support out of the box in Windows. Unless of course, you cough up for SQL Server and maybe even some client license add-ons into the bargain.
I also really, really love how every Microsoft employee has it drilled into them from an early age that any decision made, in reality for the pure benefit of Microsoft, is actually a decision made for the benefit of customers and as a result of extensive customer feedback! This is so deeply embedded in them I'm sure they believe it themselves now:
Today I have an update about how we are delivering some of the WinFS technologies. It represents a change to our original delivery strategy, but it's a change that we think that you'll like based on the feedback that we've received....It's great technology and we are super-excited to be productizing this way. And most importantly, it's what people have been asking for - as we work with customers, we're constantly hearing that they want many of the technologies to be more broadly available in the data platform products. That feedback was taken seriously."
Yer. Especially where it means more money for us.......
The fact is, what they are trying to do hasn't really been done before. AND, they're trying to tack it onto the enormous pile of legacy code that is Windows.
I wonder if the problem of integrating it into Windows itself stems from the fact that next to ZERO file formats that are currently in widespread use by the computing world know anything about "metadata", which is kind of key to the whole "SQL as a filesystem" concept.
Plus, I've always wondered how they thought all that metadata was going to get there in the first place. Most users don't even bother to name their files properly (e.g., every folder is named New Folder), and now they're expected to *decribe* them, too? Doesn't seem likely.
Then they should have designed it better, managed the project better, etc. Again, they were just demoing WinFS two weeks ago! What you're saying is that there's nothing wrong in your book with promising a major feature for almost half a decade and then going back on the promise.
"Sufferin' succotash."
The funny thing is that thier entire monopoly is based on an amazing promise... everyone knows they promised DOS to IBM before they even bought it. Unfortunately they have a long standing pattern of this sort of behavior.
But Microsoft has always done this. Every OS version they've ever released was preceded by years of promises, many of which ended up being completely false and misleading, or at the least, poorly-thought-out or overly optimistic.
Let's face it: Microsoft's best product, from a marketing point of view, has always been the version they haven't released yet.
Meanwhile, other companies, in and out of the open-source world have already delivered everything in Vista, and everything that Microsoft promised but will fail to deliver in Vista, but you know there are many, many vendors and customers who are holding out for Vista based on Microsoft's promises.
At this point, it seems Vista has been reduced to bug fixes and eye candy. Now, that's not bad, but it sure isn't worth $200 or whatever they are going to charge. Based on what you will get, it should be called XP 1.1 and cost $25. (Of course, based on what it cost to make it should cost about $5000, but that's not the customers' fault...)
You really want Vista? Install Linux or buy a Mac and then run 2000 or XP in Parallels or VMWare for those Windows apps you can't live without (there are some good ones). It will be cheaper and you can do it now.
You are in a maze of twisty little passages, all alike.
It's amazing the economy came to rely on a company so unreliable.
Microsoft isn't unreliable, not when viewed from the proper perspective. Microsoft is almost one-hundred-percent reliable when it comes to pulling the wool over the eyes of gullible customers, which they have managed to do to a customer base numbering in the hundreds of millions. That kind of reliability doesn't just happen, you know. It takes true dedication and an unwavering belief in one's own rightness. Ask yourself just how many politicians would give their left testicle to dissemble with such awe-inspiring efficiency. When someone can perform some complicated task with the appearance of effortlessness, it is a sign of true competence in action. With Microsoft, lies and deceit come so naturally one has to believe that one is in the presence of greatness.
Of course, if they'd focused even a fraction of that effort to the end of producing reliable software, we wouldn't be having this discussion. Granted, in the past several years they've improved substantially, but that still leaves untold millions of copies of Windows 3.1, '95 and '98 to be explained.
The higher the technology, the sharper that two-edged sword.
Unless you count the new start menu
How is this fundamental? Stardock's WindowBlinds has been offering the ability to create a custom start-menu for years.
the "everybody's a user" security model,
Microsoft had the ability to implement this in Windows XP. They've supported Limited User Accounts since Windows 2000. Its a change in default user settings, not an earthshaking new security model.
the sidebar
Does Google Desktop ring a bell? How about ObjectDock?
the bundle of included apps
Oh, you mean new skins for Minesweeper, Wordpad, and Solitaire? Or do you mean 3-d chess? Last I heard they weren't even including a basic office suite. For a 7-gig disc, I expect more.
Face it, Vista includes little that's especially new, even for Microsoft.
We all know what to do, but we don't know how to get re-elected once we have done it
Most of these things that you mention are fixes to sub-par elements of the OS. These aren't new innovative things to be excited about, these are basic functions that any OS would be embarrassingly incapable without, in short, the things you mention are the ante to just keep playing in the next-gen OS game:
... Windows Movie Maker? Windows Mail? You can't say that with a straight face surely!?
- The new start menu is not an enhancement, just more functionality glummed into an 11-year old UI device stretched way beyond breaking-point.
- Sleep mode is something Windows should have had half a decade or more ago, it's practically a goddamn necessity with a portable.
- "Everybody's a user" security - a huge flaw with Windows that is finally seeing some action, unfortunately looks like there's plenty of tuning to be done before it actually works.
- The sidebar - seriously, you're excited about a technology you can already have (Dashboard, Konfabulator etc.) and implemented in a boring, unimaginative and sceen-hogging way?
- Print system - I'm not qualified to comment
- Bindle of included apps - such as...
- WiFi networking which remembers the settings of each wifi network you connect to - um... come on, 6 years wait for THIS?
- "Performance Statistcits" - god, go download one of the dozens of benchmarking apps... why does this make you want to buy Vista at all?
- 64-bit support - seriously, it needs this to even be in the game, it's not some special feature to trumpet above any other OS, it's an absolutely basic necessity.
The only thing you mention which IS slightly exciting to those watching Vista is the new compositing system, Aero. Which will allow some nice effects and finally decent non-flickery, back-buffered drawing to sceen.
Talk about scraping the barrel, these things that you seem so excited about - they're nothing but the absolute basic necessity to even have the OS worth considering in 2007 when it may be released. Where are the things that make you really excited about the OS, the things that make it special? The things that elevate the experience of using the OS rather than a tick-box driven nightmare of minimum-level-of-attention-to-detail copy-cat features.
This sig has been deprecated.
Since when do people buy windows?
They'll just go along buying computers with windows pre-installed.
Of all the features that I have seen or heard of that are supposed to be in the final incarnation of vista I cannot find one that is not either 1) a tweak, 2) something already available in either Mac OS or Linux or 3) A fix of something that is broken in XP. In short Vista is not at this point coming across as innovative. Actually if I think about it it appears that its Microsoft playing catch up.
Even Aero isn't innovative, I've been playing with Sun Microsystem's Project Looking glass and whilst it doesn't do a lot for me (it makes my laptop an interesting talking point when giving presentations....) and it feels like something dumped on top of the OS to make it look like a major change.
I cannot see corporate users migrating to Vista for any real reason, even the "new" security model isn't going to be a winner there as it will break any application that through lack of proper design requires admin rights (and there are a few out there).
As for it being the biggest change since the 3.11/95 upgrade Im confused how you could even relate the two. Windows 95 was a totally different user experience from Windows 3.x. This isn't. If you look at the last real upgrades for home users (excluding DOS and whatever interface was thrown over it (buttons for DOS anyone?) it was windows 3.11 to 95 for a huge difference in usability, 95 to 98 for a massive boost to hardware support and management (in my opinion anyway) and then 98 to XP for the benefits of NT (After all I don't know many home users who got their hands on 2000 and I discount ME as it was appalling...).
I see no innovation and no reason to upgrade if you are still using Windows. As far as RAM and CPU usage, Well Im not sure I am fairly confident that you could get Vista slimmed down to normal XP performance, but then I can get XP to perform quite well, it just takes a lot of effort. Realistically though Vista is going to be on a new PC or you are going to have to upgrade something (probably add more RAM or upgrade your graphics card rather than upgrade your CPU but still.)
The really sad thing is that 6 months after the launch there will be a huge number of users, and why? because its the best OS? because its worth upgrading to? because its more secure? No. It will have a user base because it comes pre-loaded on N number of new PC's.
"For example, with FAT16, there's no way to get the file-changed notifications that Spotlight needs to know when to reindex a file."
Of course it can. The filesystem is irrelevant. A feature like this would be present in the driver's "write" function (or at a similar level), and could be implemented in several ways (which I'll skip so as to not get bogged down in details). The storage medium matters not one bit.
Certainly some of their competitors have been impacted financially. Microsoft's decade of promises, vaporware and FUD surrounding Cairo/OFS/Storage+/RFS/WinFS has been used to help persuade their customers to not migrate to other platforms. "Stay with us and you'll benefit from this wonderful new filesystem we're developing that will solve all your problems!"
Well, it turns out that most of those statements were false. Nevertheless, Microsoft retained those customers who were tempted by the promises, and other vendors lost out on the corresponding revenues.
This incoherent rambling is massively overrated and not even correct on most factual points.
How does that make you qualified to say someone who does use Microsoft products is not qualified to comment on Vista features?
Too many people I talk to (and read) have the wrong idea of what the latest WinFS iteration was and what it was supposed to do. WinFS was intended to provide sql query like filtering to files on unstructured file metadata. While this is certainly 'file search' for those who know how to make it work, it was not a consumer friendly mechanism. It was intended to be the underlying API that a consumer friendly search like desktop search would use.
If you read the article(s) you will understand that a) it grew into something more / bigger that overlaps with ORM mapping (and what was ObjectSpaces) b) it will exist in the not too distant future (next SQL server version) and c) consumer friendly desktop searches and apps you probably haven't even thought of will be able to use it. For example, why not search SNMP trap data, html log data etc. the same way you search for a file? You will be able to.
Obligatory MS bash - the fact that it has taken so many iterations does worry me as to how much longer it might take before I can actually use the technology. We are probably talking at least the next version of office - 4 years away? Thinking back, that was obvious even from the last public WinFS CTP, so I guess this whole 'WinFS is dead' (which it's not) announcement is no surprise, just a re-packaging of what I already knew.
It is not the end of the WinFS concept - it is the end of the WinFS stand alone product and the beginning of the concept's availability as an enabling technology.
slashdot troll = you make a compelling argument I do not like the implications of.
Any user or prospective customer is qualified to state opinions regarding the software, at least informed opinions at any rate.
$.02
The Christian Right is Neither (Christian nor right). See: Matthew 23, Matthew 25, Ezekiel 16:48-50
NT 3.1, 3.5, 3.51, 4, 5 (Win2000), 5.1 (WinXP), 5.2 (Win2003)(, someday Vista, likey at 5.3)
.NET, OLE, etc)
NT 3.1 through 4 did not see tremendous changes under the hood. MS removed functionality (OS2 subsystem), added the 95 GUI, and moved around a lot of the management interfaces. In NT4, they brought in OpenGL, and later, DirectX3.
Win2000 brought AD and DirectX5+, and a new driver model. WinXP brought a new GUI, and a bunch of bundled apps. Win2003 brought some GUI cleanup over 2000 Server, and some additional security, as well as extensions.
The changes become a lot smaller as you go through the products.
Given that the code base was so difficult that MS had to scrap the Longhorn tree based on XP, and start over again with the current version, the 2003 code base, it does not show well. They've patched and extended and patched again the code. They've moved piles of stuff into the kernel, and then took a lot of it back out. They've added in more and more basic APIs to do things, and they all have to be maintained. (GDI/GDI+, MFC,
Vista, so far as any of us can tell, is not a huge change. They've made a lot of UI alterations. They've changed around the bundled applications. They've added yet more APIs to maintain. They mucked around a bit more moving things back out of the kernel. As I said, just that was so hard for them to do, that they had to scrap it once and start over again.
So what's left after all this stuff has been removed? Taking account of stuff that's not available for download for other OS's (such as IE 7) and not already in XP by default in some other form (e.g. Outlook Express/Windows Mail):
- Higher system requirements
- "updated" GUI
- Windows DVD maker and some other small utilities
- Lowered privileges (welcome to the new millenium, Microsoft)
- Parental controls
- Speech recognition
- That flash/disk thing
- Drive encryption
To me, that looks like stuff that most people have available to them in one form or another and which, if they don't, they wouldn't ever use. GUI changes are ten-a-penny for Windows, DVD software is bundled with every PC that's capable of doing it or available free, parental controls are widely available for free or quite cheaply (and hardly EVER used anyway because they are just NOT reliable... I speak as a primary school technician who's had to explain that if you google for "Little Red Riding Hood", it's quite possible to find unpleasant stuff that will bypass a filter), speech recognition? Hell, my browser's done it for at least two versions, the libraries are installed with no-end of utilities and it still doesn't work very well at all - most people don't even HAVE a microphone on their computer because there's no practical use for it. There are also dozens of decent voice-recognition programs out there that tie into Windows just fine. Drive encryption is, again, easily available. However, it's not something I'd recommend a newbie to turn on until they were sure they had ten backups of their key or some kind of recovery disk.
It really bugs me that Windows, in every previous incarnation, has missed out "obvious" features, tweaks and utilities and instead bundled stuff that nobody wants/uses. Now they seem to be finally taking that cue and putting stuff that's easily available as freeware into Windows and releasing it as a new OS. A previous example would be Windows Firewall (which came along just after software firewalls had established themselves as a necessity). Unfortunately, they are still missing the obvious things that both ordinary people and the average tech NEEDS in an OS, mainly concerning control over what programs can and can't do.
A fully-installed Windows system has always needed some freeware to prop it up. Let start with "choosing which programs can run at startup". I install Startup Control Panel and StartupMonitor because I WANT to know when a program deigns itself so important that it should want to run at startup without asking me first. I get a pop-up dialog and a choice of whether to allow it to do so, EVERY time it tries. That's useful. That's simple. Complete IT-incompetents realise what it's asking and say no unless they think they need it - it instantly stops computers slowing because of accumlated startup/taskbar icons like RealPlayer, QuickTime, Adobe Acrobat etc. that DO NOT NEED to be loaded at all until I decide to load a RM, MOV or PDF (ARRGHH! What a stupid idea to "preload" these sorts of apps! Anyway...). Where was the OS facility to do that? You could regedit. You could go into some obscure menu in later versions of MS System Information (if you even KNEW how to find it, which most people don't). But nowhere did you EVER get a choice of "do you want to allow this program to run at startup?" or not.
Now, hopefully, someone who's run the Beta will tell me if that's in Vista or not. I would hope yes but I haven't heard of it yet. Also, I always install a Print-To-PDF driver of some sort (depending on the client, either freeware or something from Adobe). It turns ANY file (even Publisher) into a usable, transferrable file format that will print out on any machine (so you can transfer the PDF to a computer with any printer that uses the same paper size and it will pretty much work... that is invaluable in my line of work). Now it looks like MS finally caught up and then Adobe said no? I can't say I blame Adobe and I'm actually wondering why
To be more precise, it's the hooks into the VFS layer that allow notification for Spotlight to update. Take a look at bsd/vfs/vfs_vnops.c in xnu (you might need to sign up for a free ADC Online membership in order for that link to work). In particular, take a look at vn_close - in particular the ..._fsevent calls. (NOTE: this is not a published interface, and is subject to change without notice. Don't start using it in your apps unless you're prepared for an app to stop working in a future release; it might continue to work in future releases, or it might not.)
My home directory at work is on an NFS server, and everything under it it's indexed by Spotlight. It happens to be on an HFS+ partition on the server, but Spotlight on my machine has no clue that it happens to be on HFS+, and the indexing of my stuff there is done by Spotlight running on my machine, not on the server, so no hooks into HFS+ were used to do the change notification (because, among other things, no such hooks exist; the hooks are in the VFS layer, above the individual file systems).
BeOS had a 64-bit database-like file system way back in 1995!
"I'm not a cool person in real life, but I play one on the Internet". Galley
Can store data
Can index meta data
has an oo like storage structure
Can do transactions .Net and add meta data parsers for their own documents, done and one point more for selling .Net.
Basically what Beagle does anyway.
Most modern filesystems can at least do 3 of the four criteria, so all you need normally is an index for meta data. Be recognized that around 92, Apple recognized that upfront around 2002 when they hired the Be Guy, while Microsoft is simply stepping in Bees footprint and in the footprint of many CMS system developers, having failed with such an approach blatantly.
Microsoft could have had it cheaper, get the Lucene search engine, push it into
You seem to have neglected the closed source, and increasingly discarded use of resource forks in MacOS as an old predecessor to this type of additional data about a file, stored orthogonally to the file itself.
Managing that kind of data gets extremely nasty over a network: because a file and its resource data need to be handled in a very atomic fashion, it makes network filesystem handling far more difficult. Anyone who'd worked extensively with Appletalk could have told Microsoft this: it looks like they just wasted several years of effort and a lot of development work to verify that it's a very fragile approach.
I really don't get it - are they so out of phase with reality where:
Anybody involved in search application development will tell you that the current systems are in their infancy and that they are inadequate due to the lack of a relational structure and most importantly a query language.
WinFS was Microsoft's chance of actually being ahead with a technology. Given both the bad press that Vista has been receiving as well as the fierce competition they are facing, they really needed this.
A lot of people say they want Microsoft to get a clean start but they don't really understand what that means. Running your legacy support in a VM is a really stupid idea. Remember when Aplpe did that with Classic? It was horrible. Overnight they basically lost all compatibilty with scanners, tons of printers, camera and lots of other products. Not to mention that various production enviornments involving audio recording, video creation, or digital imaging is going to suffer for years to come.
It took Apple about 3 years to get OS X usuable. That's counting from the day 10.0 was released but in reality they had been working on OS X and NeXT for at least 5 years before Panther came out. And Apple had the NeXT code base to draw from. Microsoft would be starting from scratch so it would probably take them much longer and there's no way they could spend time creating "cool" new features because they'd be busy re-implementing basic stuff like DVD burning and file encryption.
Sure, every other Linux kernel can throw backwards compatibilty to the wind because they know everyone will upgrade quickly and all the software is free but Microsoft doesn't have that option. The Windows world would be split between the new and the old for the better part of a decade and the cost of re-buying all your proprietary software (Photoshop, Dreamweaver, Nero, Acid, Office etc.) would be prohibitably expensive.
For some people the move from Win9x to WinNT (XP) was difficult enough and that was after 6 years of transitionary work on the NT platform to make everything as seemless as possible. You guys say you want to see that future but in reality you really don't.
Plug: http://www.osreview.com/ for more commentary
Actually, you don't need Vista for that.
That may be true, I'm not the poster above. But as someone who understands WinFS intimately, Reiser did indeed miss the mark in a few ways. Hans seems like a very smart guy, I would have loved to have him on our team. It appears from his comments that he sees things from an FS viewpoint. The problems he brought up (especially his first, small files and large files) were simply not a major issue with WinFS at all.
Right or wrong, WinFS took more of a database/object store point of view. If you were to draw a spectrum from fully-schematized data in table with rows/columns/relationships all the way to a bit-bucket optimized stream store file system with no indexing, you could paint WinFS a bit closer to relational than to file system.
WinFS did expose hierarchically nested data storage (unique in its particular design to the database world as it did not use nested tables, sparse columns or other typical implementations) and in addition to that also exposed a hierarchical namespace, security, and file system compatibility (again unique to the database world). It also promoted a particular system for canonical schemas for data that is used every day by people who use computers and ON TOP of all this provided first-rate services for synchronization, full-text search, backup and restore, antivirus...
I worked on WinFS. It was extremely ambitious, and yet it did actually solve these problems. Performance is always a concern, and of course it wasn't perfect, but it really wasn't the defining characteristic or even that important of a factor in why WinFS was cut short. I won't pretend to understand the exact reasons, so maybe I'm wrong, but let's look at perhaps the most ambitious part of WinFS, which wasn't even mentioned above: the API.
WinFS took itself very seriously in trying to design an API that would work over all data, relational and file system, and there are a LOT of differing design considerations when you try to address such a large, diverse set of developers. Some care nothing but for pure access speed, some care nothing but for how easy it is to learn and use. In WinFS the only way to access the store was via the API (not via raw SQL statements), this put a lot of pressure on getting it right. Now try getting it right when you are building it from scratch and from just a small pool of users.
I'm not claiming that the API was the reason that WinFS plan changed, but consider that now the plan for WinFS is to more closely embrace its relational roots, and perhaps more narrowly target the database developer base as a beachhead for the next generation of database access APIs, and I think you can see why the WinFS web postings were so focused on playing up the future in the next version of SQL Server.
You have to look at the full breadth of the ambitious vision before you realize that cutting back this vision still leaves a whole lot that can be delivered without necessarily involving a service pack so large that it makes XP SP2 look like... well I guess XP SP1.