From: Brad Templeton (brad@main.templetons.com)
Date: Wed Oct 06 1999 - 20:26:04 CDT
On Thu, Oct 07, 1999 at 01:42:10AM +0100, Andrew Gierth wrote:
> That is, however, mostly irrelevent to the purpose of this group;
> everything in the above paragraph is a matter of policy, not technical
> standards.
My point was we've argued a lot of policy, and a lot of not that useful
stuff on technology too.
>
> Brad> Instead we've argued about how long a line can be, like it
> Brad> matters. It doesn't.
>
> It does matter if you're writing code; people are _still_ writing new
> news servers and clients, and maintaining and extending old ones, and
> as long as that is the case there is a need for technical standards.
That's not what I mean. What I mean is we all -- and I'm sure myself
included -- have to sit down and figure out what the bad consequence
is if we don't fight and trade some compromises. What it means to say
that it's not that meaningful what the exact line length limit on a
subject or references line is is that most of the proposed forms will
work.
The other thing that was needed, which I proposed many times but got
no participation in, was some surveying of opinion, so that people could
figure out which issues were split 50-50, and which ones 90-10. When
you meet in person, you get that just from being in the room or if need
be by quick show of hands. Mailing lists seem to not do that well.
However, my broader point is that for the success of USENET, many of
the questions of large debate don't matter a lot either way. It's
stagnating as it is, and fixing the number of entries in a references
line won't alter that.
I would rather have just said "no technological limits on anything" --
article size, header size, etc. And left the rest to policy. The idea
of setting limits (other than ones unanimously agreed to) in standards
expected to last years for computers is silly. It was like IBM setting
a limit of 640K of memory, which caused all sorts of trouble in later years.
Old software does *not* need to be conformant to the new standard. It either
gets conformant or dies.