Many people have found bugs in unverified parsing routines in the Linux kernel. All I'm asking is for someone to find a single one in one of Microsoft's new pa

Many people have found bugs in unverified parsing routines in the Linux kernel. All I'm asking is for someone to find a single one in one of Microsoft's new parsers. Surely, if specification bugs are just as commonplace as other kinds of programming issues, this shouldn't be too big of a request. I'm mostly just tired of people pretending that formal verification produces buggy software at a rate that is in any way comparable to "ordinary" programming. Empirically, it doesn't, and people who claim it does have scant evidence to back up those claims. Are there specification bugs? Sure. Do they happen at the same rate that bugs in comparable unverified code (let alone C code) occur? Not even close.