It seems there is a long standing belief that in IE does not pass the
event object to the handler functions attached with attachevent() to a
DOM element.
The really long standing allusion on the matter is that attachEvent is
a buggy implemented equivalent of addEventListener so everyone is
trying to analyse it from this position.
In fact attachEvent and addEventListener are what I would call (as a
joke term) "homofuncs" by analogy with homophones.
Homophones are words that are pronounced alike even if they differ in
spelling, meaning, or origin, such as "pair" and "pear".
"Homofuncs" would be then programming entities from different
environments that are producing seemly similar results in some
particular usage despite they differ in their nature and real purpose.
This post explains it in details and with samples:
<http://groups.google.com/group/comp.lang.javascript/msg/78281e500f55079c>
The explanation to the attachEvent "argument" is that it is trying to
jump on the bound method in the global context of its behavior, but
there are neither first nor second so it's just a "poor lost baby".
I don't think that we should report it to Flanagan though you are
welcome to. Over years it became a well-established tradition to talk
about badly broken attachEvent method on IE. Maybe just keep it "entre
nous"?