Danny,
If there is a need for general interoperability around a topic proposing a
new header is not necessarily the most important step.
The document is a work in progress, which is attempting to summarize a number
of discussions which were had on the filtering list several months ago.
Rather than present a finished document for discussion, the revisions are
being announced on the list to give people an opportunity to criticise it early.
Any thoughts you have on it, or on what you think general interoperability
should
mean, are worth posting to the list.
I don't get a clear and unambigous statement of requirements or
specifiaction for this functionality.
Currently, the document has had a few people adding bits in several places,
so it's a bit out of focus. One of the authors is currently streamlining
it to improve the focus, and a more readable copy should be available
in the next few days.
Please don't hesitate to post your thoughts on areas which are confusing or
seem out of touch. They may well be, but someone (you :-) needs to point it out.
Roughly speaking, there are two interrelated ideas. The first idea is
that a common header for filtering results, with key/value pairs
will allow any software to read the output of any compliant filter.
Some of the key/value pairs must have universal meanings, while other
pairs can be filter-specific. This removes the need for custom, filter specific
parsing of headers.
The other idea is that if different software ("downstream") can look at headers
added earlier in the message lifetime ("upstream"), then there are data
integrity and trust issues which must be addressed, resulting in formalized
rules such as e.g.: "write a single header for each filter, not two or three
which must be combined to be meaningful".
On Oct 21 2004, Danny Angus wrote:
I'm not convinced, from first brief reading, what the benefit of all this
is.
"X-" exists to permit non-standard headers. Most users of "X-" use them for
either conveying proprietery information or for making a human readable
trace.
If there is a need for general interoperability around a topic proposing a
new header is not necessarily the most important step.
I don't get a clear and unambigous statement of requirements or
specifiaction for this functionality.
Don't get me wrong, I haven't had time to fully digest it, but on first
reading it seems to have a very large scope and it responds with a proposal
which is not currently supported by anyone.
d.
--
Laird Breyer.
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg