Re: Filtering Header Draft Discussion

J.J.Bailey wrote:
> When reading headers, I prefer to see them concise, self-explanatory,
> and most importantly, easy for both humans and programs to parse.  The
> format used by SpamAssassin and Bogofilter seem to meet these
> requirements, so I adopted them for my own filters.  For example, I
> have a general DNSBL filter that allows users to check any DNSBL they
> wish.  Example header:
> 
> X-DNSBL: 219.146.207.114, dnsbl.njabl.org=127.0.0.9
> 
> Ditto for a general RHSBL filter:
> 
> X-RHSBL: 01o.com, rhsbl.ahbl.org=127.0.0.2
> 
> After reading the draft, I see the X-DNSBL entry looking something like:
> 
> X-DNSBL: ip="219.146.207.114"; dnsbl="dnsbl.njabl.org"; value="127.0.0.9";
> 
> I don't see any problems with making these changes.

The contents of that header look approximately like one would expect, with simple 
key-value pairs. However, part of the goal of this draft is to standardize the name of the 
header that filters output, so that readers can look for the same header name every time. 
Additionally, having complete trace information about the software adding the header is 
desirable.

In other words, a mail scanner that added a Filtered header indicating DNSBL results could 
look something like:
Filtered: name="DNSBL-Check"; version="1.23"; host="mail.example.com"; function="dnsbl";
     dnsbl-name="NJABL"; dnsbl-server="dnsbl.njabl.org"; auth-received="Fri, 5 Mar 2004
     19:50:45 -0800 (PST)"; ip="219.146.207.114"; response="127.0.0.9"; tag="open-relay";

> Does anyone have a feel for how much effort is being put into client-
> side filtering?  I'm curious to know what's driving the standard and
> who needs what where.

At the present, I can't really speak to that question.

Phil Miller