Re: 8.6

From: John Stanley (stanley@peak.org)
Date: Wed Feb 04 2004 - 21:06:38 CST


 Charles Lindsey (chl@clerew.man.ac.uk):

>Since the ultimate purpose of the Netnews protocols is to cause articles
>written in one place to be displayed in lots of other places,

That is not true. The unltimate purpose is to transport articles from one
place to many. There is no requirement that they wind up being displayed
anywhere. For example, when people distribute binaries via news, they use
news protocols, but the don't actually display the articles when they
arrive, they run them through something that repacks them into the
original binary, which may or may not be displayable. For another example,
I believe there are ncar groups that use news protocols to transport met
data from place to place (and if there are not, then someone else probably
is).

>However, there is another issue which is exceptionally important for the
>proper operation of Usenet (and to a lesser extent for email also) and
>that is Followups. The proper way to do a followup is most certainly a
>protocol issue (because Usenet will not work as intended without those
>rules),

For the limited extent that USENET "will not work" without some rules for
followups, we have the References header, which is defined for that
purpose and is mandatory in said followups.

>and this in spite of the fact that a user can manually override
>almost everything that the rules say a followup agent is supposed to do.

In my experience with news, it is much much much much more common for a
user to override the Subject header than it is for him to override the
References header. That is not to say that I've not done it myself, but
I did so with full knowledge of the standard and because I truly was
starting a new thread.

>Agents that read (and display) the
>outgoing article are entitled to expect that the article they see was
>constructed according to the protocol, and if they get screwed up because
>the followup agent screwed up, then they are entitled to shout
>"interoperability".

I have NEVER found a followup agent that did ANYTHING wrong with an
article where I modified the Subject to remove the "Re: " it wanted
to put there. Not once. Not in all the years I've posted.

If a followup agent breaks when someone removes the "Re: " hack, then
that agent is BROKEN, and once again I'll say, it is not our job to worry
about such badly broken software. This is NOT interoperability.

>...and some that thread
>using References also provide for separating out sub-threads where the
>Subject has changed.

Except they don't know that it is a "sub thread", all they know is that
the Subject content of THIS article is not the same as the content of
THAT article. I cannot recall the number of times I've seen Subject
content changes in the form of "Boo is good for you" to "Boo is not good
for you", and it is the same thread, just that one poster has decided to
make his opinion clear in the Subject.




This archive was generated by hypermail 2.1.7.