Com-Central.net

It would appear that the first release won't now be until the 2nd quarter of 2007 and general release will follow that.

Microsoft and Gartner are squabbling again, this time after the analyst predicted yet another delay to Microsoft's already late-running Windows Vista operating system.

According to Gartner, Windows Vista is unlikely to ship before the second quarter of 2007, pegging it for launch in April next year, with "broad availability" not expected until at least the second quarter. For good measure, the analyst firm noted that Microsoft "consistently misses target dates for major operating system releases". Microsoft must be thanking Gartner for that little reminder.

You know, I realize that being late means less revenue in the short run (missing the christmas shopping spree), but I've never seen the big deal with extending launch dates. Would you not rather have a completed product later than a buggy and useless one sooner? Granted, a release date that changes more than 3 times is annoying. A company should be realistic about this stuff from the beginning.

Vista will still be buggy even released late. LOL

Software testers have a set time to get the bugs out of any software (this is a professional requirement).
They have to match 1) a release deadline with 2) the amount of time required to test the product thoroughly, compared to 3) an "acceptable" risk to the reputation of the company from releasing a product too early.
In almost every example to date, M$ have proved that they just don't care about the last part of that equation.
The only reason they're extending the release is to achieve some level of functionality.
Most of us accept that there will be an update patch with Windows releases, and I for one will be waiting at least until Service Pack 1 until I dip me toes in 'Lake Vista.

Have to agree with Xcaliber here. This certainly doesn't seem to fit the profile of 'not caring about an "acceptable" risk to the reputation of the company from releasing a product too early'.

"For good measure, the analyst firm noted that Microsoft "consistently misses target dates for major operating system releases".

Cripes, that's stated like it's a BAD thing......

Gartner Inc. is an analyst/consulting firm, Suits and bean counters. Check their "products and services" for a great example of managment-speak pap. Pbbbt. "Optimization, Decision and Pricing Solutions" which Joe Blow could care less about...riding the coattails of developers who actually create the thing that does the actual work for the end user.

In any development effort slightly more complex than just the classic "Hello World!" program, a balancing act always occurs over when to stick a fork in it and call it a release, versus recognizing and fixing the "show stoppers". No developer team is ever really finished in their own minds, and you can go on for eons polishing and refining away as well as fixing every small little thing and polishing all the door handles....'creeping elegance'...before it even sees the light of day. Which in complex applications basically comes down to the date "never"!

And by definition, testers don't get the bugs out, they find them - software engineers do the hunting and killing. Yes of course there is a schedule with milestones that everyone follows and bases their own goals on, schedule slip is the prime indicator that the vendor/developer DOES recognize and care about a quality product over schedule and perceived maximized profit. I see schedule slip as a victory for the development team in the trenches! Geez, the two work hand in hand - a quality product will do more to maximize your long-term profit than junk shipped on a date that coincides with whatever's on the Gantt chart, MS realizes this as much as anyone, they're not exactly newbies to the O/S market. Just because there's a proper schedule laid out so people know what to do and when, doesn't make anyone a bad guy.

These days it's all about "process" and how it's managed. In many Software Development Life Cycle models, the developers are farther removed from the actual testing, and the actual test steps and such must follow a written test plan as opposed to good old ad-hoc beat-it-up testing by the folks who wrote/know the thing and how to really break it. If the test plan isn't written to expose a weakness, the weakness goes undiscovered and schedules slip when they're finally found by accident (i.e., good old "ad hoc" in BETA). But at least the test plan and process was followed, thank god, and the analysts who document, sell and promote the process which incurs inevitable delay get to harp on the developers for not being perfect.....

Anyone asleep yet??

I agree with most of what you said.
Though current schedule slips can't be a victory for Micro$haft developers.
Nor, for that matter, their testers.
It's slipping because of the bugs the beta testers found.
Great chunks of it work, but not together.
They're having to re-write about 60% (figures provided by the contributors from M$'s development team, though anonymously).

There are basic qualifications for software testers anow where one of the first things they learn is the "Release date/Test time/ Acceptable risk" factor now. They are taught that no software house tests its software 100% anymore (they're taught this so the individuals don't get too stressed out about not meeting set targets when their time runs out). It's accepted industry wide.

You and I are the final testers nowadays.