Oblig XKCD. Most recent one in fact.

Performed the same test. Received similar results. Unless your ISP's DNS presents you with issues (e.g. redirects, slow response time, downtime), it's probably the fastest option. If you are concerned with privacy issues, then your ISP is once again your best choice, since they already know what sites you visit.

Microsoft had a huge beta program for Windows 7, and those are the versions that Microsoft wants people to test and run. The biggest problem is that internal builds, which were never meant to go public, sometimes make it out, because a lot of people aren't making the correct distinction between released builds and questionable leaked builds. They could include alternate approaches that alter APIs, scaring the hell out of partners, new (possibly major) bugs that haven't been caught by QA, or simply be incomplete or old. When people see these, and think that this reflective of the quality or direction of the software, that's a big problem for a guy in marketing that tries to communicate the quality and direction of the software.

Let's be honest here - if you want to play around with NeatOpenSourceProject, are you going to go to sourceforge and download the newest daily build, or are you going to go download the latest release, regardless of how much older it may be?

That's the same thing I do to your Mom! Zing!

Doing formal software verification on any non-trivial sized code quickly devolves into a nightmare. Proofs like this have long been the goal of the software verification community, and it's good to see this positive result. Regarding the verification, it's actually quite useful. If you run into a bug, it's because you're using the kernel incorrectly or the kernel has been improperly specified. This verification eliminates the need to search for bugs within the kernel itself. Its behaviour has been verified as exactly meeting the specification.

Haskell is a functional language, not a logic based language. Prolog is the most common logic based language. You use logic statements that describe known facts, then give relations between facts. To get a solution, you provide a query, and Prolog performs an evaluation using your facts and relations. Neat stuff.

I would hope he's being serious, because that statement is correct. Correctness of software is defined as meeting the spec. Does this mean the spec is right? No, but software that correctly meets the spec is doing exactly what the software was specified to do. A bug appears when the software doesn't meet the spec.

Only if the damn neighborhood kids forget to shovel the walk.

Having worked in sales for a small software company, i can speak first hand as to the volume of complaints and requests that come in from customers. Talk to any customer for more than 10 minutes and you have a dozen feature requests and complaints, half of which are unique to that customer. In open source, the response "do it yourself" is quite suitable to deal with oddball concerns, but obviously that can't work in a closed source environment. The problem isn't that companies ignore customers. It's that so many requests come in, it's just plain impossible to satisfy the great majority of them. To appease the greater number possible, companies simply must address the most common concerns first, and only address smaller ones when it fits into their development plans and future goals. I remember meetings where sales and tech support would sit down with the dev team and we'd prioritize every existing request, bug, and fix. From lists hundreds of items long, we'd end up with 10 or 20 things above the line, and everything else would be ignored by necessity.