From: John Stanley (stanley@peak.org)
Date: Sat Dec 21 2002 - 23:12:32 CST
David Barr (barr@visi.com):
>Several have noted that many past implementations
>have assumed it was case sensitive.
Yes, the argument that we must do it this way because some implementations
have done it that way. Shall we remove all things from this draft that are
different from how things are currently done, or is your argument specific
to this one thing where you don't agree with me?
> It's also been noted that the field
> is a newsgroup-name, and that newsgroup-name is case sensitive.
While the field may be labelled "newsgroup-name", the entry "poster" is
most certainly not a newsgroup name, and since it is not, there is no
argument that it must be case-sensitive. I've already dismissed the
results of this argument -- that it is just too hard to make the test for
"poster" case-insensitive, and the implication that there might be a
newsgroup-name "Poster" or "pOster" someday. Do you have something new?
>And "Because John says so" does?
I'm not saying it has to be case-sensitive. I'm saying that there is NO
REASON for this artificial limitation, that the arguments about it being
too hard to do a case-insensitive test are just bullshit, and that adding
this item now -- long after last last last call -- needs serious
justification. No, "because Henry says so" isn't sufficient, and I've gone
a long way beyond "because I say so".
Matt Curtin (cmcurtin@interhack.net):
>I'm not a single person ...
Oh, are you counting yourself as two or three today?
>but I still
>came up with a good reason, that being that widely-deployed software
>makes that assumption.
I think Henry's way of answering that would be to say that we are not
responsible for the incorrect assumptions that broken software makes.
My way of answering that is that 1) apparently not all of it does, thus
the reports of "Poster" not being accepted AND our new permission for
agents to accept it case-insensitively, and 2) this draft says lots of
things that are different than the way things are currently done. I've
already dealt with what you thought was a good reason.
>...which would make it possible for someone to write a
>completely USEFOR-compliant client that will create articles that INN
>rejects.
Not if INN is USEFOR compliant, and if it isn't, it is broken.
>(As already noted, I think case sensitivity in the header sucks, but
>we don't have the option of creating a standards-track document that
>ignores some of the most widely-deployed software in use.)
Your position is silly. Of course we have the option of creating a
standard that changes how things are done. That's exactly what I was told
when I tried using the "group charter is to document current practice"
argument against a change, and I've seen nothing to change the status of
this group since. If we can create other things, then fixing this is
certainly within the scope of our activity. Since INN will have to be
changed to be USEFOR compliant for the other things, changing one test for
the word "poster" to case-insensitive is nothing.
If you truly think case-sensitivity sucks, then the time to speak up is
now. The standard is being written now. I suppose it sucks worse to agree
with someone you don't like, though, so I understand you arguing in favor
of something you think sucks.