Now Windows Longhorn is long gone, witness reflects on Microsoft's OS belly-flop
- Reference: 1743079814
- News link: https://www.theregister.co.uk/2025/03/27/looking_back_at_windows_longhorn/
- Source link:
It's been just over 20 years since the [1]Longhorn project , as initially envisioned, was reset. Plummer worked on the project – it was one of his last at Microsoft before retirement. "Longhorn is a tale of big ambitions, technical nightmares, and an extraordinary comeback," he [2]said .
We'd argue that many users who were content with XP but frustrated by Windows Vista would take issue with that last point.
[3]
"Microsoft wanted Longhorn to reimagine what Windows could be," Plummer said. This meant a completely new user experience, some fundamental changes behind the scenes, and, of course, WinFS (short for Windows Future Storage).
[4]
[5]
Plummer described WinFS as one of the "crown jewels" of the new operating system. Rather than the file-based approach used up to that point, WinFS would utilize a database, meaning users could search for files by content tags and relationships rather than by name and folder.
It was, Plummer noted, "a bold plan."
Unhelpful Microsoft help denies helpless millions help [6]FROM THE ARCHIVES
There was also a shift to managed code with .NET, a new presentation subsystem, and a new framework for communications and web services. "Windows XP's colorful Luna interface was going to be yesterday's news," said Plummer. "Longhorn would be sleek, modern, and visually rich."
"It's not often that an operating system promises to reinvent multiple core pillars all at once. Longhorn was aiming for the sky but – and you knew a 'but' was coming – ambitions often come with big challenges."
[7]
Plummer recalled the storm clouds gathering as development ramped. Microsoft staffers connected to the project could see it buckling under its own weight.
"One major problem," he said, "was that with so many new features being developed simultaneously, the Windows codebase became fragile and bloated." WinFS was a major culprit, but pretty much the entire stack was tottering. "The integration of all these pieces became a nightmare.
"We had a tradition at Microsoft called 'Dog Fooding,' meaning using our own daily builds of the OS to do our actual work. For Longhorn, dog fooding was tough. Builds were often too unstable to run for very long.
[8]
"The degree to which eating dog food is tolerable can be measured in direct proportion to how good the dog food actually is. And this was not good dog food."
It's a familiar story. Developers would check in code, not realizing that their changes would break something else. Getting a stable build where everything worked together without major bugs proved increasingly difficult. Plus, there was the need for backward compatibility.
"The Longhorn team was walking a tightrope," recalled Plummer. "Innovate aggressively, but don't break Windows. Not an easy balancing act."
Then came the infamous [9]memo in 2002, in which Microsoft's head honcho, Bill Gates, declared security to be the company's top priority. Securing Windows was the right call, but it meant that development on Longhorn slowed as engineers, including Plummer, were called upon to patch holes in Windows XP.
Even with teams scattered, Plummer said Longhorn feature creep continued. "It's demoralizing as an engineer when you can see the light at the end of the tunnel getting further away and not closer."
Around this time, Plummer took a sabbatical from Microsoft and decided to retire. He said the Longhorn fiasco played only a small part, but "it almost seems like I dodged a bullet."
[10]Ex-Microsoft engineer resurrects PDP-11 from junkyard parts
[11]Microsoft whiz dishes the dirt on the Blue Screen Of Death's colorful past
[12]The port of the Windows 95 Start Menu was not all it seemed
[13]A cheeky intern nearly turned MS-DOS into NSFW-DOS
"I'm an incrementalist, not a revolutionary," he said. "I prefer to add things one solid feature at a time rather than reinventing everything all at once as part of some grand vision."
The fate of Longhorn is well documented. While the consumer team continued to add features, taking the attitude, according to Plummer, that "Consumer Windows could be less rigorously engineered than Server Windows," the server team took a methodical approach to ensure stability.
It all came to a head when, as Plummer recounts, Dave Cutler – who had led the development of Windows NT – noted the severity of the situation and suggested that the codebase used for Longhorn be switched to the one the server team had been working on.
"It's never pleasant to consider throwing away years of work, and it certainly bruised some egos to admit that the Longhorn project had gone too far off track," said Plummer.
The death knell came with the reset of August 2004 and the public announcement that the existing Longhorn codebase would be scrapped. Instead, the Windows Server 2003 SP1 codebase would be used, with the most important, or most complete, Longhorn features added.
Plummer retired before the course correction, but said the announcement was shocking. "Imagine telling hundreds of developers and testers, 'All that work that you've done for the past three years, we're going to set that aside and do something else.' It was devastating, yet oddly enough, a relief for some.
"Devastating because nobody wants to see their hard work tossed out, but a relief because, by this point, everybody knew it wasn't working. The project had become a death march of sorts."
The result was [14]Windows Vista . The operating system retained some of Longhorn's features, such as translucent window effects, but discarded others, including WinFS.
"Was the Longhorn project a failure?" mused Plummer. "Well... yes and no."
"It failed to deliver on its original promises. WinFS was never shipped, many of the more radical ideas were cut, and the schedule slipped dramatically. It was also a management failure as it took a near-crisis for Microsoft to course-correct.
"But Longhorn's legacy is complex and, in many ways, a positive one. The reset and recovery from Longhorn shaped Microsoft's culture and engineering practices going forward."
Plummer regards the story of Longhorn as a cautionary tale. "It shows that even a company with virtually limitless resources and a track record of shipping big products can get in over its head.
"The story of Longhorn is one of hubris and humility; Microsoft reached for the stars, stumbled, but then humbled itself to regroup and ultimately deliver something worthwhile." ®
Get our [15]Tech Resources
[1] https://www.theregister.com/2004/08/27/microsoft_decouples_longhorn/
[2] https://youtu.be/RpRZ8BQiiMo
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/oses&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2Z-WEMB3ezlDjyunEIgjVjQAAAAE&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/oses&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Z-WEMB3ezlDjyunEIgjVjQAAAAE&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/oses&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Z-WEMB3ezlDjyunEIgjVjQAAAAE&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[6] https://www.theregister.com/2007/03/08/msdn_gloom/
[7] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/oses&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Z-WEMB3ezlDjyunEIgjVjQAAAAE&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[8] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/oses&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Z-WEMB3ezlDjyunEIgjVjQAAAAE&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[9] https://www.theregister.com/2002/01/17/ms_highest_priority_must/
[10] https://www.theregister.com/2024/08/23/build_your_own_pdp11/
[11] https://www.theregister.com/2024/08/02/who_wrote_windows_bsod/
[12] https://www.theregister.com/2024/07/29/windows_start_menu_port/
[13] https://www.theregister.com/2024/04/08/ms_dos_easter_egg/
[14] https://www.theregister.com/2006/03/27/windows_vista_reader_survey/
[15] https://whitepapers.theregister.com/
Re: WinFS = Document Management System not invented here…
It's not whether they could, it's whether they should.
Folders and files are a FAR better way to go than anything else that has been conceived of. A file system is not a database, a database is not a file system.
It's just stupidity.
A company with virtually limitless resources ... can get in over its head
Unfortunately for Microsoft, they can't really claim there wasn't a historical precedent - and it does sound an awful lot like they were treating [1]The Mythical Man-Month as an instruction manual. But even IBM was unable to avoid [2]history repeating itself . And Apple had its own [3]development nightmare that ran on in various guises for around 8 years.
Interestingly, all of these projects involved a re-imagining, to a greater or lesser extent, of storage models. I doubt that's because the existing models are intrinsically better, but the concepts are not only firmly embedded in existing software, but also in users' (and developers') consciousness.
[1] https://en.wikipedia.org/wiki/The_Mythical_Man-Month
[2] https://en.wikipedia.org/wiki/IBM_Future_Systems_project
[3] https://en.wikipedia.org/wiki/Copland_(operating_system)
Re: A company with virtually limitless resources ... can get in over its head
Each of those companies decided they knew better than the customer and the customer would just have to learn to like their products or lump it.
We don't call them tech douche bros for nothing.
Re: A company with virtually limitless resources ... can get in over its head
No, it's because the existing models ARE intrinsically better.
Storage models DO NOT need to be re-imagined. ALL attempts to do so just make everything worse.
Deliver something worthwhile?
Oh how I laughed at that quote.
M$ hasn't delivered anything worthwhile in decades.
Re: Deliver something worthwhile?
and they have been around for decades ;)
Re: Deliver something worthwhile?
When XP was getting close to EOL, I said that if there was nothing better by then, I'd switch to Linux. A year after EOL, I finally did; been running Ubuntu ever since. I keep asking why I didn't do it sooner.
There still isn't anything better than XP, just bigger, more bloated, and using more resources. Oh, and getting a new paint job every couple years in the name of "progress".
Re: Deliver something worthwhile?
"There still isn't anything better than XP, just bigger, more bloated, and using more resources."
Yeah, that's the point. There's this upgrade cycle that consists of faster and more powerful hardware, but nobody needs it, because Windoze runs good enough on current hardware. So they break the operating system & do it again only worse.
I'm completely out of the "new computer" market, 100%, because I switched to Mint and it truly does not need "new hardware".
I've been rummaging around the e-scrap bin down in the parkade for the last couple few weeks and have so far resuscitated 6 discarded laptops, all of which run Mint like a son of a belch.
"Pay" for a computer? That concept is obsolete!
Re: Deliver something worthwhile?
Well Vista certainly wasn't it.
Windows 1 0 is not an entire failure, but 11 is certainly going in that direction.
The thing, XP (SP3) is certainly the last Windows OS that considered that it was "your" PC it was running on.
Everything since considers that it is Borkzilla's PC, and Borkzilla can do as it pleases.
I can't wait for the day when I can retire, erase all that shit and install Mint and once again have a "Personal" computer.
Re: Deliver something worthwhile?
I'm not sure about Vista since I never used it, but I do remember Windows 7 was the first one for me that declared you must be running a pirated copy if your computer was airgapped and it hadn't phoned home in a while.
Re: Deliver something worthwhile?
I don't recall that ever happening.
I mean, I only EVER ran pirated copies, so that's probably why. I'd even wipe new computers and install pirated copies, because they were cleaner than the bloatware that Micro$hit and OEMs shipped.
Re: Deliver something worthwhile?
Did you derive some pleasure from the popup when you log in that says 'you may be the victim of counterfitting'?
"Ah.... 'victim', you say?"
Re: Deliver something worthwhile?
The thing, XP (SP3) is certainly the last Windows OS that considered that it was "your" PC it was running on.
AFAICR XP was the one that introduced the idea of phoning home. I decided that by that time Linux was good enough to replace SCO which dual-booted with Windows on my previous PC and become by sole OS.
Re: Deliver something worthwhile?
I think it phoned home, but it didn't get particularly bolshy if it wasn't networked and couldn't.
When was it ever otherwise?
Windoze has only ever succeeded by hanging onto the coat-tails of DR-DOS and VMS.
This is the only occasion they tried to write an OS themselves - what did you think would happen? ;)
Re: When was it ever otherwise?
Back around '86 or '87 I talked to a prof at UW who'd done some consulting over in Redmond. "What's it like there?" I asked. "A lot of smart people wandering around the halls reinventing the wheel" he replied.
But not smart enough, if history be the judge.
Deeper problems
One of the other problems with Longhorn - and subsequently Vista - was complex technology with no well-defined purpose.
I was an industry analyst at the time (no, not Gartner) and recall an advisory meeting with the marketing team where they showed me (among other things) the translucent windows feature and then asked "can you suggest any ways we can explain to customers how this is useful?"
This was a point where the launch for Vista was just weeks out. Apparently, at no point in the long slog of Longhorn had anybody stopped to ask "wait, why are we doing translucent windows?".
And this problem pervaded Longhorn. There did not seem to be any grounding in solving users' problems or improving their experience. So when things started to go off track, they had no basis for deciding what they should ditch, what they should fight to keep, and what could maybe be tackled another away. The technology became a goal in itself rather than a means to an end.
Re: Deeper problems
Yep. Apple was infected with the same nonsense.
One of the first things I do when setting up a new copy of macOS is setting "reduce transparency". And there are lots of other little tweaks that have to happen now, like making all windows NOT vanish when clicking on the desktop, disabling tiling, turning the status bar back on in Finder windows, fixing the backasswards mouse scrolling (which has also infected Windoze and Linux now), and turning icons in the title bar of Finder windows on, among others.
Change for change's sake, not because it actually improves anything, only makes things worse.
... the Windows codebase became fragile and bloated ...
That's new, how?
The only **truly** good OS's that MS ever produced were DOS 3.11, Windows 2000 And Windows 7.
It is beyond me why they didn't just keep plodding along with W2000.
It was a solid base that could have been perfected then expanded.
W10 and W11 are abominations unto retupmoC and should never have seen the light of day.
As for .NET, if there is a hell, then there is a very deep pit reserved just for its developers and proponents.
Nope.
They have NEVER released a good OS. The best they've ever accomplished is "barely functional" and Stockholm syndroming you into thinking that was "good".
It's not quite *that* bad. Windows 2000 was reasonably stable and functional. Windows XP SR2 and Windows 7 were pretty stable too. Windows 10 Pro is usable. 95, 98, ME, Vista, and 8 were junkyards dressed up as a Jenga tower and about as solid. Windows 11 is basically 10 with lots of additional enshittification. Since there's a huge application ecosystem built on Windows, it's not bad to have one of the stable Windows releases on the desktop, if you know how to stabilize it. OOTB it's pretty bad, though. I've avoided 11 and don't know if it is really possible to stabilize it to the usable Windows 10 Pro level. (The Home versions are a bad joke.)
I could never stand the washed out look of XP. It was possible to get it looking a lot like 2000, which made it usable, but only just.
I don't consider W10 to be usable at all. And W11, as you say is just another version of W10.
If ReactOS is ever finished, I will be one of the users.
Until then, I will stay with W7.
Washed out? The default XP colour scheme was lurid- alongside the skeumorphic shading, it was reminiscent of plastic childrens' toys.
Ironically, it was miles better if you chose the restful green colour scheme, and even silver was somewhat better.
But it's the worst colour scheme of the three that XP will be forever associated with, purely because it was the default.
(While I'm here, I was never a fan of that lurid, oversaturated, overrated hill background that ageing millennials are now apparently getting nostalgic for. Urgh.)
And here we are, 20 years later, Windows Search is still broken, and Explorer can barely show thumbnails of popular formats, let alone search through document content.
Have you ever noticed the people just 'learn' to work around the search box on the start menu without thinking?
Quickly type 'notepad' into it before it's had a chance to wake up and it'll display a bunch of help results or web search results for the word 'notepad'. It's then second nature to just delete the 'd' off the end and, hey presto, Windows has worked out you might want a program it has called 'notepad'.
I swear it's been doing that for over 5 years... yet they think it's time to pile resources into co-pilot.
"Longhorn is a tale of big ambitions, technical nightmares, and an extraordinary comeback,"
Nowadays the mantra seems to be "two out of three ain't bad."
Nah. It's "one out of three ain't bad". And the one is always "technical nightmares".
WinFS = Document Management System not invented here…
The stupidity of WinFs was that it’s intent could so easily have been achieved using one of the several DMS’s that inserted themselves into the file access of MS Office.
I note that even today 365 does not include any document management functionality. Just shows how little things have really progressed in 30 years…(*)
(*) document management / saving files in a user friendly way, was seen as an issue with Win3/WfWg.