From: John Stanley (stanley@peak.org)
Date: Mon Dec 16 2002 - 15:10:17 CST
Henry Spencer (henry@spsystems.net):
>> I guess not rejecting articles where an agent has treated the value as
>> case-insensitive isn't a benefit, huh?
> No, it's not.
An odd definition of benefit, I expect.
> Why would an agent do that?
From the new section 6.7,
| Although the keyword "poster" is case-sensitive, followup agents MAY
| choose to regognize case insensitive forms such as "Poster".
Why may they choose to do this if they will never see it? Because maybe it
will be appearing on the wire and isn't that important a thing to waste
time worrying about whether the P is p or P or the o is O?
Because the person posting has entered the value in a version reasonable
for humans, and there is absolutely no reason for it not to be valid,
other than "Henry Spencer said so"?
>There is no benefit in accepting articles where a misbehaving agent has
>generated/tolerated misspelling "poster" as "psoter", "postur", or
>"Poster".
"Poster" is not a misspelling, it is a perfectly reasonable capitalization
of the word "poster". There is no reason for the agent to be labelled
"misbehaving" simply because you have defined this capital P as a problem,
when the fact is that there is no technical reason for it to be a problem.
>No, this is a new report, of a new mistake.
Who cares if you've seen the report before or not? It is a report of a
problem in existing code, which makes it an issue for how this draft
documents existing behaviour.
>But the mere fact that someone
>recently made a mistake does not demand that we bless it as existing
>practice.
The mere fact that nobody has complained to you, personally, of this
problem before the one report that was distributed here does not make this
mistake only recent.
>Since the very earliest days of Usenet, people have been known to misspell
>newsgroup names.
We are talking about neither a newsgroup name nor a misspelling. Please
stay on-topic.
> Because much existing software does not accept it,
There is stuff in this draft that existing software will not accept. Are
we prohibited from doing anything new because of that?
> ...son-of-1036 explicitly says ...
Son-of-1036 is not a standard. What that document says is historical
curiosity, nothing else. Had it gone through the standards process, then
we'd expect it to have some weight.
> and it would be a new special case requiring special handling
Oh, please. If you argue that doing a case-insensitive string comparison
is some "special handling", then you've as much as admitted that you've no
real reason not to allow "Poster".
> (nothing else that can appear in Followup-To is case-insensitive).
So what? You've already got to do a comparision to "poster", all you need
to do is use the case-insensitive version for that one comparison. You do
know that most standard libraries that deal with strings have a
case-insensitive comparison, don't you? You use strcasecmp instead of
strcmp. You put /i at the end of the perl regexp. That it isn't any more
"special handling" to do the comparison using that version instead of some
other version.
The only reason to require a case-sensitive "poster" is if you really,
honestly, truly expect there to be a newsgroup named "Poster" someday.
Otherwise, this limitation exists only to stroke egos and prove that you
can demand things that have no technical reason for being.