ietf-mta-filters
[Top] [All Lists]

Re: values of notify ":priority" tag

2005-12-09 15:48:43


What are the use cases for priorities other than those in RFC 1327?
Sure, infinite priority gradations sound appealing in theory, but what
will they really be used for?

The fact that we ended up with Priority/X-Priority/X-MSMail-Priority makes me
think that the needs of the original field proved insufficent, and so further
fields were defined.

Perhaps, but it doesn't follow that the insufficiency was a lack of possible
values. As Kristin pointed out, there are various quite different semantics
associated with these fields. X-Priority in particular seems mostly to
be aligned with the X.400 concept of importance as opposed to priority, and
I've seen X-MSMail-Priority used in several different ways even among
Microsoft software.

Personally I hate the fact that the task list in my iPAQ
only allows me only three levels.  There seem to be no technical restrictions
for allowing arbitary levels, so why make such a crippled feature?

This is also incorrect, I'm afraid. There may be little difficulty associated
with using lots of levels when all it does is order things, but that's not all
these fields are used for. The X.400 concept of priority, for example, affects
both processing priority and message timeouts. (Actually, this last may be more
a GOSIP thing than a pure X.400 thing - I no longer recall which of the many
specifications says what about it.) Each priority level ends up having a bunch
of settings associated with it, and that means there's real overhead associated
with having lots of levels.

Personally, I wonder why priorities are considered so useful and the
option is not just dropped, but that's me.

I think the notify extension should focus on being a framework to be
filled with methods that interface different media, inside and outside
the internet.

Michael

I agree with this approach.  Priority notifications can be added later if
they prove desirable enough, but I think the feature is going to dramatically
slow down the spec, needlessly so.

I also have no problem with not having a general priority/importance parameter.
If we do have such a parameter we have to get it right, and that means the
reality of there being a limited number of levels available in lots of
environments has to be accomodated.

                                Ned