On the surface this appears to be a bigger problem than it is.
Microsoft has identified five areas that have to be ready before Vista can ship: reliability, security, performance, driver coverage and application compatibility.
While I cannot be certain, it does make some sense why we are not seeing more progress from vendors in the area of driver coverage and application compatibility, because of the way Microsoft builds to spec. Meaning very simply, that vendors can't deliver on drivers and updated apps until Microsoft itself does more work to finalize the new OS in the areas of reliability, security and performance - which must come first.
After the first three are completed, then drivers and apps are forked back into the build.
Vendors know this and are likely much further along than any one of us expects, or can really know - and just as huge leaps "seemed" to progress forward between Vista BETA 2 and 5536/RC1-5600, we'll see [in the RTM build], what appear to be similar large leaps forward.
What appear to be huge gaps in many areas, followed by eqaully large steps forward, are likely as much to do with "how" Vista is being built than they are a lack of vendor progress.
That leaves a big question regarding reliability, security, and performance - just how close is close... and perhaps as big a question, how ahead are vendors and how solid are their assurances to Microsoft that drivers and apps will be ready when the tangs are brought back into alignment? My bet is that we are seeing, perhaps for the first and last time, inside the mind of Allchin himself, and how a really large piece of software is built. Like ourselves, Microsoft engineers can't eat the entire buffet at once, either.
We recently built a suite of software this way - and it scared the crap out of the customer as progress "seemed" to be stalled over all, but was instead, right on track and when all was forked back into the build, it again, "seemed" to jump forward in huge leaps. I have to say, it was much harder on senior staff and myself - trying to keep what the end products would look like clearly defined. However, by building smaller, bullet proof components, we were able to ensure that the final product was ready to go and as close to perfect as engineers can build software with available tools.
|