Between 1993 and 2001 Microsoft released a new version of Windows every year, with the exception of one year: 1997. That's nine major releases in ten years.
The concern back then wasn't that Microsoft wasn't shipping enough major OS releases, but that it was shipping too many!
It was only with the wrap-up of the Antitrust case in this decade, that Microsoft's product delivery began to dry up.
2000 found Chairman Bill talking up the successor to "Whistler" (which became Windows XP), for 2002 delivery.
As we pick up the trail a year later, Blackcomb had slipped correspondingly.
Summer 2001: the Blackcomb blockbuster may be late
"Whistler" had been regarded as a code cleanup and To-Do features that didn't get into Windows 2000. By summer 2001, the extent of the ambitions for its successor became apparent. It involved a relational datastore at the heart of the OS, with the NTFS file system as a plug-in.
By the fall, Gates confirmed that the Big Bang wouldn't explode until 2003.
"2003 is the next major milestone for us in terms of the Windows release," said Gates. "That will be a very important release, a lot going on in the peer-to-peer, a lot going on in the advanced presentation environment there," Chairman Bill told the 2001 PDC.
So at that stage, Microsoft envisaged the all-singing, all-dancing upgrade "Blackcomb" to be released in 2003 or 2004, with what we called a coffer-filling release, "Longhorn", to emerge first in late 2002 or early 2003.
Spring 2002: There will be no sequel
The possibility of an "XP Second Edition" for early 2003 was floated amongst Microsoft partners, but this turned out to be a trial balloon, and the option was rejected.
Instead, Allchin said a consolidated service pack would be released later that year. This turned into SP1, and was indeed released on time in September 2002.
Summer 2002: It's Bill, and I'm in charge now
We didn't have to wait long for further hints. Gates intimated to Fortune's Brent Schendler that he himself had personally taken charge of Longhorn, imbued it with his "vision", and suggested Longhorn would be his monument to humanity. (Here Gates first used his "many moon shots" analogy.)
Which meant it would be very late indeed.
While not making much semantic sense, this remains the most accurate prediction made in the five year period: "A radically new version of Windows, code-named Longhorn, which, if all goes well, will come out sometime after 2005."
(We think he meant "radically different" - which has turned out not to be the case, but at least the date was correct).
In Fall 2002, Microsoft extended support for Windows 2000 into 2003, a sure sign that XP's successor was going to appear soon.
Spring 2003: ... And this is what our new file store would look like, if we had written it
In March, Longhorn leaks, and where the new relational datastore WinFS, should be, there's a UI of what the new relational datastore will look like.
Developers are instead thrilled by the new "Sidebar".
In May, Microsoft confirms that that the relational file store WinFS won't appear in Longhorn, but sometime later in the decade. Or, whenever.
Fall 2003: Technology dictates our roadmaps. Really
In October 2003, Bill Gates floated the possibility further delays, suggesting "Longhorn could be 2005 or 2006."
"This release is going to be driven by technology, not by a release date," said Bill.
And that was the last time the big technology revolution was ever talked about.
By December, Brian Valentine was heading the "core OS group" and Gates big dreams were postponed. Longhorn would appear in 2006, even if it meant dispensing with features - the biggest, WinFS, having already been thrown out. Features began to be jettisoned.
Spring 2004: There will be no Shorthorn - official
By now the delays, and missing code, were hard to cover. Microsoft was again obliged to explain why there would be no interim release, this time dubbed "Shorthorn" rather than "XP SE".
Last July Microsoft finally christened the baby.
And the "SideBar" returned.
Bootnote: A Terry's Chocolate Orange then to Gartner's Tom Bittman for suggesting back in 2003 that Microsoft might the miss the must-ship target of 2006. Ridiculed at the time, they were right. Sort of. ®