Talking about DNS and reverse DNS the other day at lesson, one of the guys asked me about what reverse DNS is good for. In fact, the only useful example I could come up with, was the reverse resolution of the SMTP HELO command used to avoid easy spammers. Is there anything else it is really useful for? I bet it is, but I can’t think for what!
Another thing. Discussions about incredibly-expressive grammars and fast parsers is always on. E.g., we have context-free grammars, but LL(1), LR(1), LALR(1) parsers (and so on) which manage to rapidly parse them. Ok, but: is it my impression or the cost of parsing is a minor factor in the overall equation of compilation, and most of the time is passed into optimization, parsing tree traversals, bytecode interpretation, and so on? In that case, why we don’t use more powerful parsers (like Earley-likes, whose modifications run almost always in linear time too and can parse any context-free grammar - even ambiguous ones - without any modification)? I suppose it’s worthier to write grammars easily and correctly rather than trying to adapt that C grammar to LALR(1) or stuff to gain that small percentage of time…
Posted by mattia as dns, grammars, smtp, spam at 10:45 PM CEST
