Re:

From: Seth Breidbart (sethb@panix.com)
Date: Mon Mar 01 2004 - 10:42:54 CST


John Stanley <stanley@peak.org> wrote:
> Seth Breidbart (sethb@panix.com):
>
>>> So we can specify that because that's required for interoperability with
>>> current software.
>
>>So is handling Re: correctly, so why do you have a problem with that?
>
> Because it is not true. Any reading agent that breaks because it does
> not find "Re: " at the start of the Subject content, or because it finds
> "Re: Re" ", or because it finds "Sv: " is defective. The content is
> unstructured, which makes "Re: Re: Re: Sv: Re: Re[234]: " just as valid
> as a subject as "Re: ".

The choices aren't just "breaks" or "doesn't break", they include
"does the right thing" and "does something considered brain-damaged by
98% of all Usenet readers". Sorting "Re: a problem" miles away from
"a problem" doesn't break the software, but does annoy the reader.

The choice is apparently between a spec that is technically correct,
minimal, elegant, and not particularly useful, versus one that
specifies doing what the world considers the right thing. Since I
want Usenet to be viable and useful for the readers and posters, I
prefer the latter. There are other, better, venues than this one for
showing off my ability to write elegant definitions.

> J. B. Moreno (planb@newsreaders.com):
>
>>Turning:
>
>>Subject: I just graduated!
>
>>into
>
>>Subject: Re[2]: Antw: SV: Re: Re[3]: Antw: SV: AW: antwr: antwort: re: re:
>>raf: si: sv: Antw: Re[2]: SV: Antwr: Si: Sv: Antwort: Ref: Si: Sv: Aw: Aw:
>>Sv: Re: I just graduated!
>
>>seriously impacts the usefulness of the Subject.
>
> So use a followup agent that doesn't do that.

So why are you opposed to specifying that followup agents shouldn't do
that (or individual steps of that)? Or do you believe that some
people _do_ want such behavior and we need to cater to them?

Seth




This archive was generated by hypermail 2.1.7.