> Stepping back: trying to engineer a new mechanism in a standards > committee makes me nervous (as we have generally been taught that this > is a Bad Idea - witness pthreads). Just what problem are we trying to > solve here?
Sorry to (partially) answer my own question, but I just remembered the answer: if we don't do something new, then we have to standardize the numbers - is it S10network or S20network? I still might prefer this (closer to existing practice, simpler, &c), but I haven't gone out and compared a bunch of distributions, or thought too much about which entries need to be standardized (network and portmap at least).
