Like Carlos I had always felt that metaservers should be neutral, but I
have changed.  The change may not be permanent, but the results so far
have been spectacular.

It is difficult to endorse or punish via information flows, since there
is not a single metaserver.  The metaserver I'm running has changed
policy.  The policy is only effective while there are few players.  The
policy has limited long term impact, because:

1.  clients can be written to use both metaservers and merge the
results,

(the COW code on Linux does that with my UDP metaserver query mode,
added in 1999, which still hasn't hit the Windows client code base ...
I'm sure the current Windows clients can follow the trend),

2.  servers now respond to metaserver queries, and clients can be
configured to query them directly,

(again, code contributed by me, this time in February 2006, currently in
COW on Linux and the Vanilla source),

3.  players can learn how to avoid the restrictions, especially with
enthusiasts begging them on continuum to learn how to type "netrek -h
something", and spamming the message board with pro-sturgeon political
statements, or advertisements for clients,

4.  in the same way that players know to circumvent bans by changing IP
addresses, server owners know very well that they could change their
hostname or IP address to circumvent the policy.  (I was actually
encouraged by some to turn off the UDP solicit mode, te he).

Regarding the player count, or robots reported as players problem ...
before we opened the metaservers to allow anyone to list a server, we
were able to enforce some unwritten policies ... such as non-offensive
hostnames, servers that actually can be logged into, etc.  With the UDP
solicitation, we lack any of those controls, and so I feel a touch more
justified in exerting some control on "my" metaserver.

Regarding previous policy, a minor observation ... when RSA client
binary verification was first implemented, the metaserver followed suite
by reporting whether a server was compiled with RSA.  The metaserver
didn't restrict what servers were returned.  Now we learn that the R
flag doesn't identify whether the server actually demands verification
at the time.  Patches welcome.

-- 
James Cameron    mailto:quozl at us.netrek.org     http://quozl.netrek.org/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://mailman.us.netrek.org/pipermail/netrek-dev/attachments/20061125/a7fc3fe9/attachment-0001.pgp