Re: Implementing NodeList

Re locality of "liveness" -- no, it doesn't help those who haven't seen the
spec before; we don't know whether absence of a comment means it doesn't
apply or just that the authors forgot to mention it this time. If you
really mean for there to be a difference, I do suggest either different
types (see previous note) or explicitly saying "this one is live though
others aren't" _and_ "this one is not live though others are". Or both.

(Different types strikes me as the best answer. Liveness or lack thereof is
an artifact of which call was issued, but is architecturally part of the
behavior of the NodeList... and no class should have two conflicting
behaviors without clear indication of which is applicable. You could add a
boolean attribute to the beast to indicate which case is active, but
subclassing is arguably a more OO way of representing that distinction.)

Received on Monday, 27 July 1998 20:51:04 UTC