Re: Summary of Issues Raised to date

On 16 Mar 2004 at 18:17, Yakov Shafranovich wrote:

> > I hate to be the one who suggests a lowest-common-denominator 
> > approach to a problem, but I really think that's what we need to be
> > doing here. Instead of trying to solve *all* the header-related problems
> > we can think of, why not start off with the simplest thing we think we
> > might actually be able to achieve and build on that?
> 
> The questions is - what should be included in such header. And this is
> where all the problems start :) Can we perhaps try to narrow down the
> question of what should be included in such header and what role such
> header would play?

Let's look at this in terms of what the header is actually going to be used 
for: I would say that in (considerably) more than 90% of cases, the 
existence of the header is all people will be interested in - they'll have a 
rule along the lines of

   "If there is an 'X-This-Is-Spam' header, junk the message"

Many MUA filtering engines are reasonably simplistic, and only those 
that are truly scripted are going to be able to parse arbitrary values 
within the body of a header line anyway.

The only people who are going to be interested in more information than 
this are people who are interested in it as feedback for tuning their 
spam detection systems, and in cases like that, the information that will 
be of interest to them is going to be system-specific. Given the wide 
variety of ways in which content evaluation systems work, I doubt it will 
be possible to standardize this information in a way that will cover all 
cases, nor that there is even a particular value in that kind of 
standardization.

I think that if we could come up with a simple header name that we 
could all agree to use in our systems to indicate that a message was 
spam, we'd have achieved something. There would be nothing to 
prevent us from adding other headers as well as the dictates of our 
systems required, but just agreeing on one header that Joe Public could 
use reliably for spam / ham differentiation would have value in itself.

> A side point this leads into is a more general question of MTA/MUA 
> interaction. Some people would really like to have an ability to have a
> "this is spam" button in the MUA which will feed back to the MTA. Or the
> MTA letting the MUA know that some of the previous messages delivered
> should be considered spam.

I think this is probably outside the ambit of this group, isn't it? I mean, 
it's an idea that may be worth pursuing, but wouldn't it be better to stick 
to the aims in hand? Alternatively, since plenty of people may want to 
discuss this type of thing (myself included), perhaps we could decide on 
the basics first then enlarge the discussion to include topics like this 
after that's done?

Cheers!

-- David --

------------------ David Harris -+- Pegasus Mail ----------------------
  Box 5451, Dunedin, New Zealand | e-mail: [email protected]
           Phone: +64 3 453-6880 | Fax: +64 3 453-6612

Explanatory note on a British Government form from the 1960s:
   "Separate departments on the same premises are treated as
   separate premises for this purpose where separate branches
   of work which are commonly carried on as separate business
   in separate premises are carried on in separate departments
   of the same premises."