What is really causing sort of the rewrite on Longhorn? Is it these sort of demands from the partners, is it the technological difficulties of the project, or is it even personnel and execution issues at Microsoft?
There's no rewrite going on here. WinFS, I'd be the first to say, is very ambitious. Nobody has ever brought together the world of documents, media and structured information in giving you one simple set of verbs that lets you richly find, move around and replicate those things.
Ever since about 15 months ago, when we moved Peter Spiro of our database group to take charge of WinFS, we've made very good progress. What happened here is, as we looked at the new things we wanted to add to WinFS, that would have only been consistent with an 07 schedule -- adding the tabular stuff and figuring out a server plan.
So we definitely were faced with a decision that Jim, Peter, Steve (Ballmer) and I were having a lot of dialogue over these last couple of weeks. What was the right thing? Was it to take Longhorn as a whole and get these super-cool additional WinFS features in, knowing that that would push the release out into '07, or was it to come up with a plan that was a bit more clever and really not give up much?
The plan we have does give up WinFS shipping with Longhorn. And so if you want my basic assessment here, the glass is three-quarters full.
The WinFS team, in terms of its progress and performance, is doing very, very good work, but it couldn't take the additional features and make an 06 schedule. That's professional engineering on its part, to stand up and say, "no, if you want us to have those features, we're an 07 deliverable".
It seems as if software is taking longer to bring to market. SP2 kind of grew in scope. Things like Yukon and Whidbey have taken more time. Has software just gotten more complicated to write? What, if anything, does Microsoft need to do as a company to reflect that reality?
Our scheduling and predictability on this project has been better than it was on OS 360 (the mainframe operating system created by IBM). So software has not gotten more complex. Software with this kind of scope of features and compatibility has always been complex. That's the business we're in.
The ongoing dialogue we always have with our customers, ISVs and OEMs (original equipment manufacturers) is always one of two kinds. One is (that) we have a date-driven release. Things that make that date get in. For MSN, most of what it does is date-driven releases. It does very regular releases because of the nature of its market, and it doesn't have the compatibility challenge.
With the operating system, customers want releases to be on the order of, I'd say, two to three years. Hardware exploitation, media, security and wireless expectations mean that you've got to have that release. But it's enough of a challenge, in terms of deployment, testing and other things, that you wouldn't want two major operating-system releases in less than (a) two-year period.
Talkback
well, he's got an honet face
That's why we let him get away with it
This is a window of opportunity people. Finally Linux does everything Windows does (and more). There is no reason not to make the change. Get out of Windows before you get locked straight back in again.
The possible reasons unanswered:
1. Mr Gates and his men were informed of too many security holes in current Windows' code by the feedback reports from Government Security Program, or other source-sharing activities. Until they complete patching those holes, or even rewrite the codes to make Windows a "Secure OS," Microsoft cannot make any contract with governments worldwide. They don't have time for blushing WinFS up.
2. Mr Gates and his men found that WinFS is too valuable to bundle with a desktop OS like XP. They might think WinFS has a possibility to be another cashcow for the company, especially in the context of Web-services.
3. As a file system, WinFS is too complicate to manage. When the storage sytem goes down, recovery of the contents will be more difficult work and takes longer than NTFS. It doesn't provide users with benefit.
4. The post-XP should be a 64-bit OS. This means that all source codes need to be recompiled with secured new compiler, and to be tested thoroughly. I cannot imagine how long does it take.
I will not buy that for a dollar.