"Jonas Sicking" <[EMAIL PROTECTED]>
There are *a lot* of features about XHR that I would like people to use in
the end. onreadystatechange would be nice to get people off of entierly
for example since having separate onload, onprogress and onerror is a lot
easier to use.
I think onprogress is useful and can be added without effecting anything
else - ie an an author won't change their code for implementations without
an onprogress, the readyState=4 will still exist where the we're done is
finished.
onload and onerror though aren't so useful, because all scripts will still
have to treat the readyState=4 situation, so there's little benefit to
authors to have them defined here, it would just lead to more duplicated
code.
2. Better for new implementations.
I can agree to that.
Well, it means new implementations have to implement more stuff, despite the
fact that authors will be continuing to author to the old format. I'd like
to see changes that can't realistically be used deferred to a later
specification, where the new neater methods are combined with new
functionality.
Cheers,
Jim.