IBM goes slow on XP update - News - ZDNet Who can blame them? My methodology has always been to watch the history of updates from a given source. After a couple of years of updates that do more good than harm you trust them enough to apply them almost immediately. I say almost because you at least try it on a few test machines first, but this can be done within hours of the updates availability. What you do after not one, but several bad updates, those that break key components and fix problems that you don't even have, or claim to fix problems but actually don't, that is another matter. Having been burned on an update from any company, I put the company "on probation" and apply new updates only after extensive testing. This is a simple intuitive approach, that I think most users apply to both software updates as well as new product releases, and even hardware purchases. I have my own shit-list of hardware vendors as I suspect most people do. I just don't buy from these companies until they do something to change their reputation. Most never do. It is a true rarity for a company to suddenly discover the value of quality products to their future business. Companies either get it, or they don't. Microsoft has never got it, other than the original release of Windows NT which appears to have been thoroughly tested. From more recent experience this must have been a fluke, or, since NT was such a departure from previous version of Windows they were just uncharacteristically carefull.