Shoes and debugging
28 Jan 2005 
in the wee hours 
Matt Winckler
I intensly dislike purchasing new shoes. Every three years or so, it comes time to replace the all-terrain footgear, and I must embark on perilous expeditions to the dark corners of Consumerdom in search of the perfect pair.
My wife and I hold differing philosophies on shoes. Her philosophy is that she would rather buy four $5 pairs a year than, say, a $60 pair every three years. Comfort, says she, is secondary; cheapness and new-shoe-ness is all. I do not subscribe to this motto, because my bonus bone would not like it.
Therefore I found myself yesterday carrying out reconaissance in places various and divers. I conducted sorties in PayLess (great name, that), Gart Sports, Target, the mall (lapse of reason), Sportsman’s Warehouse, and Factory Brand Shoes. At long last, I settled on a pair at Factory Brand Shoes that fit the bill: they were on sale, they were relatively comfortable, and they were essentially combat/construction boots disguised as business casual dress shoes. I purchased them. May they save me hassle for another three years.
Whatever wondrous works are ahead of them, however, they do not seem to be very helpful for debugging Delphi code. I have encountered that most beastly of problems, the problem of a pointer gone amiss somewhere in a hundred thousand lines of code (not necessarily mine, which makes the search all the more fun) and mangling with addresses not meant to be mangled with. The symptoms show up in the sorts of places I shop for shoes: various and divers. Nobody online seems to have any grand insight on easy ways to squash such a bug, so it looks like I’m going to have two options before me: slog through a large pile of tedious code trying to spot a wayward pointer with my naked and bloodshot eyes, or frob random code elsewhere until whatever’s trying to read the corrupted memory goes somewhere else instead. Time will tell what course of action wins out…