The release date for beta 2 of Windows XP has slipped another week, to March 21st, according to Paul Thurrott of WinInfo. The latest delay however doesn't seem to be because of any major 'gotchas' - the XP beta code has been pretty solid for some months now, so mightn't the delay be more about the people Microsoft will be allowing to access the code soon, rather than about bugs as such?

By all accounts the look, feel and performance of the beta code hasn't changed significantly in the three or four most recent builds of WinXP beta code that have been released externally. The latest of these, build 2446, was according to Microsoft released specifically for hardware testing and bug-fixing, and as the internal build 2454 that escaped at the end of last week did so via (and what an appropropriate word that is*) a major chipset outfit, its purpose was probably more of the same.

The reality is almost certainly that Microsoft already has the base code it intends to ship as beta 2, and that it's marketing rather than development that's holding it up. If you check back to the grand unveiling of XP at Seattle's Experience Music Project on February 13th, you run across a statement that might now be inoperative. Microsoft then intended to ship beta 2 within a couple of weeks, but only to a fairly narrow set of technical testers. Instead of that we got the 2446 interim 'this is not beta 2' build, which seems to have gone out to the full range of testers.

Then, the idea was that beta 2 would go out to the narrow set of testers first, then this code would form the basis for the widespread public beta Microsoft simultaneously started advertising. The two now seem much more likely to be run fairly closely together (note that licence changes in anticipation of the public beta were made in 2446), or even as just one big public beta 2/RC1, the point of which will be to achieve greater market momentum rather than to shake out problems in the software.

Which makes the widespread beta, the "Windows XP Preview Program," pretty much shipping code, with all that entails in terms of driver support. Obviously if you're going to have hundreds of thousands of people getting the code, you want to make sure even the newbies have a positive experience, right?

Whatever, it's now tricky to see how Microsoft can fit a genuine Release Candidate 1 into the schedule, and as for RC2, well... An email concerning the previous delay (to March 14th) leaked to CRN last month set down a schedule for this, and it really doesn't work now. Says the email: "When we hit RC1, we have a solid eight-week program to RTM... The RC2 release is a short milestone where we can fix any final big things."

Eight weeks ahead of RTM puts RC1 at the beginning of April, which gives ten days, maybe two weeks, between beta 2 and RC1, and of course there's a preview program to slip in there somewhere as well. The fact that this can happen without RTM slipping as well does kind of make you suspect that the code out there already is more or less RC1 anyway, which is more or less RTM. This is not beta programs as we know and love them. ®

* It has been suggested to us that the chipset outfit that goofed over 2454 was indeed the one beginning with V. Meanwhile although we've also had it claimed to us that the escaped 2454 code is fake, someone claiming to be Dennis DaMenaCe himself hotly disputes this in a discussion thread on neowin.net. We at The Reg are inclined to agree with Mr DaMenaCe on this one. We're told that build 2454 was released internally from Microsoft's main build lab at 4.53pm on March 6th.

Related stories:
WinInfo story
More WinXP code leaks onto Web, with pix
Cracks appear for latest WinXP protection tech
Early death for 98 as MS pushes WinXP - public beta confirmed