1. 4

Submitted previously, but there were no comments.


  2. 4

    I’m not going to take the time to do so now, but I believe it’d be possible to lay out a persuasive argument that Postel’s maxim as applied to network protocol design has imposed costs of a similar scale to what Hoare claims for null references. If you deem security against hostile parties on your network an important element of robustness, the net effect of the principle is really the opposite of what its name suggests.

    That said, it probably did make it much faster (and initially less expensive) to get massive numbers of unrelated machines to interoperate at “internet scale” and I don’t want to ignore the value of that while staring down the billions of dollars in nasty security problems it has led to.