From: Charles Lindsey (chl@clw.cs.man.ac.uk)
Date: Mon Nov 08 1999 - 05:22:37 CST
In <87ln8cu8v6.fsf@erlenstar.demon.co.uk> Andrew Gierth <andrew@erlenstar.demon.co.uk> writes:
>>@@BEGIN NCM HEADERS
>>Version: 0.9
>>Issuer: annihilator-1@erlenstar.demon.co.uk
>>Type: spam
>>Action: hide
>>Notice-ID: binnew.19991105173145
>>Hierarchies: alt,alt.binaries
>>Count: 15
>>@@BEGIN NCM BODY
>># intnet.net 207.90.1.201 Rk7JmPf/EEWeDEcsKiEK0A== 21.21/1
>><8e75.b03a.35c@one> alt.binaries.erotica
>> alt.binaries.erotica.breasts
>># intnet.net 207.90.1.201 Rk7JmPf/EEWeDEcsKiEK0A== 21.21/1
>><8e75.b040.19a@one> alt.binaries.erotica.butts
>> alt.binaries.erotica.sex.in.the.morning
>># intnet.net 207.90.1.201 Rk7JmPf/EEWeDEcsKiEK0A== 21.21/1
>><8e75.b045.2a8@one> alt.binaries.pictures.erotica
>> alt.binaries.pictures.erotica.amateur
>[rest of entries snipped]
>>@@END NCM BODY
>>
OK, I have has a look at the Nocem syntax. Observations:
It is possible to invent new NOCEM headers (which then apply to all the
message ids listed). So to that extent it is extensible.
It is not possible to extend the syntax of individual message id lines,
except by adding complicated comments as in Andrew's example. So to that
extent it is not extensible. BTW, what is the purpose of including the
list of groups affected with each message id? Surely that information is
readily available in the history file, at least for such of the groups as
you take?
However, I think if we were going to invent a blockcancel, then we should
seriously look at that syntax for the body. However, the question remains
whether such a block cancel needs to be a control message at all, or
whether just encouraging greater use of nocem-on-spool as it already
exists is sufficient. The only real difference AFAICS is that, if it were
a control message, it would be filed under a pseudo group such as
control.nocem (which would actually be a disadvantage for suck feeds).
Bear in mind that any new 'blockcancel' feature will be useless until it
is widely implemented. Bear in mind also that the chief aim should be to
reduce the total volume of cancels/nocems that is currently being
propagated round the net, consistent with achieving the desired effect.
>The "Hierarchies" header is non-standard and not used by clients
>AFAIK; it's there so that I can test my batching code that atttempts
>to group entries together by hierarchy so that I can support different
>NoCeM groups for different hierarchies (e.g. there's a de.* group for
>NoCeMs).
So would it be a good idea to creat uk.net.nocem? And is creating nocem
groups in individual hierarchies better/worse that using Distributions for
the same purpose, as Brad suggests?
-- Charles H. Lindsey ---------At Home, doing my own thing------------------------ Email: chl@clw.cs.man.ac.uk Web: http://www.cs.man.ac.uk/~chl Voice/Fax: +44 161 437 4506 Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K. PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5