From: Thorfinn (thorfinn@tertius.net.au)
Date: Mon Aug 02 2004 - 21:47:50 CDT
On Mon 02 Aug 2004 at 09:49:25AM -0300, in <200408021501.i72F1kfi000333@jefferson.patriot.net>,
"Shmuel (Seymour J.) Metz" <Shmuel+gen@patriot.net> wrote:
> Thorfinn <thorfinn@tertius.net.au> said:
> > What quicker phasein? The situation remains identical - no software
> > out there parses the things, and no software out there will parse the
> > things, whether we say "MUST accept" or not.
> No, the problem does not remain identical, because putting that
> language in serves as a notice to correctly recognize the presence of
> MIME parameters and ignore them, as opposed to interpreting them as
> something other than MIME parameters.
Erm, no, it doesn't. That would be "MUST parse". "MUST accept", as
I've already explained a number of times, simply means "don't crash and
burn if you see a MIME parameter".
We would be better served by adding a NOTE that future RFCs may well
decide that MIME parameters in headers are a good idea, and that
software should start leaning in that direction.
> >The same amount of software needs to change if a mythical future RFC
> >desires such a thing, whether we say "MUST accept" or not.
> No, as I've already explained.
I don't agree with your explanation. "Accept" is a very broad term, and
does *not* require any sort of parsing. And you *do* have to do parsing
in order to detect the existence of a MIME parameter.
> >That mythical future RFC is going to have to deal with having a flag
> >day, regardless of whether we say "MUST accept" or not.
> Not unless it requires specific processing as a result of the new MIME
> parameter.
This whole discussion is way off in "predicting the future" land. We
have no idea what future RFCs might want to do or not want to do with
MIME parameters, and for my money, I'd prefer we left the problem up to
them, since they presumably will know what they're trying to achieve
with said MIME parameters, where we, back here in the past, have
absolutely no idea.
> >"MUST accept" is a completely null operation.
> No. It implies changing the code to not break if there is a MIME
> parameter.
*Nothing* currently "breaks." They just don't "do the right thing",
whatever that's supposed to be.
> >"MUST accept" is just plain meaningless,
> No. See above.
Yes. See above and below.
> >when any implementor can *and will* simply say, "right, I already
> >accept the things, even though I don't parse them correctly".
> If he parses them incorrectly then he's noncompliant. If he simply
> ignores them then that is perfectly acceptable behavior.
No. If we said "MUST parse", then he would be noncompliant. "MUST
accept", simply means must *accept*, i.e., don't crash and burn.
Nothing crashes and burns at the moment. They just don't do the "right
thing". Doing "the right thing" requires *parsing*, at least sufficient
parsing to detect that there *is* a MIME parameter, and then removing
it (if that *is* the right thing, since it's arguable that with a
parameter, the header is different to one without, so just removing it
may *not* be "the right thing" at all), before going on and doing
whatever else is considered normal behaviour for that header.
And, as numerous people have already pointed out, doing sufficient
parsing to remove MIME parameters *is* actually trickier than it seems,
and *again*, so far, nobody's actually seriously argued that we really
actually *want* MIME parameters on any headers.
Do you? If you do want MIME parameters on something, then please, say
so. So far, you've said quite the opposite, assuming I'm reading you
correctly.
We don't want MIME parameters on existing headers, let's say so, and
leave it at that. There is *no* point in second guessing future RFC
writers, who may or may not decide that MIME parameters on existing
headers are a good idea or not.
And even if we were to try to second guess future RFC writers, I think
it's likely that none of them will *ever* decide that MIME parameters on
existing headers are a good idea, precisely because declaring a flag day
on a system that's as distributed and anarchic as Netnews is about as
effective as King Canute's ordering back the tides was.
Anyway, I think at this point, it's pretty clear what the general
consensus on this matter is. The only holdout is you, and you're
holding out a solution for a situation that nobody else believes is ever
going to occur.
There's pretty much no point me saying any more on this subject, so I'll
leave it at that.
Later,
Thorf
--
<a href="http://tertius.net.au/~thorfinn">thorfinn@tertius.net.au</a>
~/ For those who've come across the seas, ~/ We've boundless plains to share, ~/
~/ With courage let us all combine, ~/ To Advance Australia Fair. ~/
-- Part of the second verse of the Australian National Anthem.