Windows Longhorn slips again, becomes megaproject
Gates holds forth on Microsoft's next Big One
Yes folks it's spring in Seattle at last, the moss has gone green, and the Redmond marketing team is busy altering perceptions by dispensing prestige exclusive executive access to prestige publications. The second of the week so far is Bill himself, in Fortune, painting a rosy (albeit staggeringly vague) picture of Longhorn. Who's next? Steve? But everybody gets Steve...
Anyways, immediately the obvious bottom line of Bill's disclosures to Fortune writer Brent Schlender is that Longhorn, once upon a time the intermediate point release on the way from Windows XP to Blackcomb, the real big one, is now not-a-point-release, but is instead "a radically new version of Windows, code-named Longhorn, which, if all goes well, will come out sometime after 2005."
Much about Microsoft's wondrous OS roadmapping procedure beggars belief, but this one raises the bar (as they say in Redmond) on beggary. Unless Brent has got the date down wrong we're talking about the next major rev shipping in 2006 earliest, which given commercial realities would mean a rollout in Q2-Q3 2006, which means a gap of getting on for five years from XP to Longhorn, which means we're likely to have not one but two point releases between now and then.
That of course is judging on past experience, and we should maybe take into account that sales channels and distribution will continue to change radically over the next few years. So actually, the death of the packaged OS may occur in the interim, and Microsoft may find new ways to satisfy its need for continuing revenue enhancement through continuing innovation during that period. But it does kind of look like we've knocked Longhorn back in order to turn it into Blackcomb, and that's still weird.
Historically, the big Microsoft OS projects that take a long time and overrun have been heavily influenced by the Bill factor. Bill wants them to be great, keeps coming up with stuff that has to go in, they get bigger and bigger and take longer and longer. As we indicated earlier there's actually not a lot of meat in the Longhorn information Gates has given to Fortune, but Bill is now chief software architect, driving Longhorn development, and from what he has to say about what Longhorn has to be and do, it's clear that the OS as currently envisaged falls into the 'lots of stuff, major project' category. Actually, according to Bill, it's the equivalent of "many moon shots." Regular readers will recall space images spilling out from previous major Microsoft projects, and may wonder if maybe it wouldn't be cheaper for Bill to buy NASA.
Says Schlender: "Bill and his teams are starting with a clean sheet of paper, rethinking what a computer operating system actually is, from the way documents and other data are stored and shared to the way people interact with the machine.... Gates' geeks are completely overhauling the operating system [so] they'll also have to redesign most of the company's other software products and services to take full advantage, including the MSN online service, its server applications, and especially Microsoft Office."
We've heard the 'complete overhaul' stuff before of course, but Microsoft's plans to put a database at the core of the OS file system, originally in Blackcomb but then in Longhorn instead, mean that in this case it has to be true. Having bitten that particular bullet, all of the other stuff Bill talks about becomes obvious and necessary. So yes indeed, the applications line has to be rebuilt, and yes indeed, MSN morphs once more.
The decision to go with all of the major stuff in Longhorn was taken last August, at the prompting of Steve Ballmer, according to Gates. Prior to this the plan had been to make changes incrementally, but Ballmer apparently voted for a big bang, with all of the components synchronised, instead.
Gates's list of what's planned for Longhorn is largely user's eye view, classic eye-candy of the sort that gets bolted on to the company's interim releases, but given that we're currently talking about a major overhaul, these ought to be more integral to the finished product than has often been the case in the past. Gates alludes to the database angle by asking of current operating systems: "Why are my document files stored one way, my contacts another way, and my e-mail and instant-messaging buddy list still another, and why aren't they related to my calendar or to one another and easy to search en masse?"
Why, indeed? We know the answer to these questions, so does Bill, and they are not particularly flattering. Still, if Microsoft is genuinely trying to design a decent OS that fixes these problems, however belatedly, then surely we can call it progress.
He also suggests using the computer for screening phone calls and emails, getting in touch with you when you're out of the office, letting you pick up your data from anywhere, with any device, arranging meetings for you, making it easy to set up web sites that can do group scheduling, finding and reading documents, etc etc. All this is laudable stuff if properly executed, and Microsoft has plenty of first class coders who could execute it properly. But it really all depends on how well the core OS is executed, otherwise we're back to the bolt-ons and go-fasters.
Gates has had a list of "ten key Longhorn scenarios" he has had devised, and regrettably only six of them, People, Annotation, Real Time Communications, Storage, Authentication and Security, and New Look, have made it to the pages of Fortune. Each of these categories has however had a developer team assigned to it, which could be bad news, depending on how good or bad the coordination is. Further detail on what Microsoft actually means by each of these scenarios will also be necessary before one can judge their validity.
On the upside though, Gates is bashing a prime directive, which is "Where's my stuff?" He accepts that the file space on a PC right now is "a cesspool," and that you ought to be able to get at everything simply and easily. If this "No 1" question is driven right through the development process then Longhorn ought to stand some chance of being useful.
Can it be done? For all sorts of reasons, Microsoft has repeatedly failed to build a great operating system in the past, even when it did go into particular projects with first rate developers and the intention of doing so. When it has come close (NT springs to mind), it has subsequently tended to undermine the product by gluing more bits on (for the sake of eye-candy, commercial imperatives, compatibility, all three) while losing sight of the basic design criteria. And while it's been doing this it's contrived to derail quite a few potentially great operating systems from rival companies.
Furthermore, the Bill Gates who currently puts himself forward as being focussed on making Longhorn technically great is the very same Bill Gates who (as revealed in numerous trial emails) has spent a great deal of time getting extraneous bits shoved into products to produce speed-bumps for the opposition. So we're not hopeful. Still, it could happen, and even if it doesn't it'll keep them busy for five years. ®
MS to micro-manage your computer
MS DRM OS, retagged 'secure OS' to ship with Longhorn?
XP successor Longhorn goes SQL, P2P - Microsoft leaks
MS poised to switch Windows file systems with Blackcomb