if a person has two moderator logins (this is quite possible), then a bot could quickly generate lots of articles which are 'cross-rated' between the two moderators. Similarly, any circular dependency will do it. Unless there are no safeguards against similar abuse, a moderator could 'skyrocket' his own rating artificially. some sort of time limit (ie. only X moderated posts per day limit for a rookie moderator) could prevent at least certain types of 'old boys network' abuse. What do you think?
are users rewarded for submitting stories? If not then this could further 'differentiate' users from flamers, a user with a good story-index will think twice before posting noise.
every moderator has weak spots. Eg. a technical guru might flame people who oppose abortion. If moderators are rated 'per-topic' (News, Microsoft, Linux, etc.), this could further increase the accuracy of the 'Slashdot Filter Machine', because after some time it 'learns' about the strengths/weaknesses of a moderator. This also might prevent good (technical) moderators from aquiring too much power to subjectively strike certain subjects.
another way to fight flames is to make it possible for a user to correct flames. A corrected flame is a reply that retracts statements made in the flame. This has to be approved by a moderator. This is a bit more than self-regulation, because it also gives mechanics to correct bad (human) behavior.
one danger might be that very active moderators make a few 'silly' mistakes just due to their activity. (high activity brings more posts, and bad days just happen, the moderator might flame someone) So maybe every 'strike' should have a 'timeout', ie. if that moderator submits useful posts afterwards, the 'mistake' gets corrected automatically.
one more rule that could be used: if a moderator 'lifts' a low-score user's post who later on becomes a moderator, then this moderator should be rewarded. (because he predicted correctly that a low-score user made a useful post) The priority should not only be to filter 'good stuff' from 'noise', but also to reward people who are good filters and dont mind to sift through filth to find the gold.
Slashdot Mirror
if a person has two moderator logins (this is quite possible), then a bot could quickly generate lots of articles which are 'cross-rated' between the two moderators. Similarly, any circular dependency will do it. Unless there are no safeguards against similar abuse, a moderator could 'skyrocket' his own rating artificially. some sort of time limit (ie. only X moderated posts per day limit for a rookie moderator) could prevent at least certain types of 'old boys network' abuse. What do you think?
are users rewarded for submitting stories? If not
then this could further 'differentiate' users from
flamers, a user with a good story-index will think
twice before posting noise.
DOWNLOAD Slackware 4.0 here
(LinuxToday appears to be slashdotted, at least from here)
every moderator has weak spots. Eg. a technical
guru might flame people who oppose abortion. If
moderators are rated 'per-topic' (News, Microsoft, Linux, etc.), this could
further increase the accuracy of the 'Slashdot
Filter Machine', because after some time it
'learns' about the strengths/weaknesses of a
moderator. This also might prevent good (technical) moderators
from aquiring too much power to subjectively strike
certain subjects.
another way to fight flames is to make it possible
for a user to correct flames. A corrected flame
is a reply that retracts statements made in the
flame. This has to be approved by a moderator.
This is a bit more than self-regulation, because
it also gives mechanics to correct bad (human)
behavior.
one danger might be that very active moderators make
a few 'silly' mistakes just due to their activity.
(high activity brings more posts, and bad days
just happen, the moderator might flame someone)
So maybe every 'strike' should have a 'timeout',
ie. if that moderator submits useful posts afterwards,
the 'mistake' gets corrected automatically.
one more rule that could be used: if a moderator 'lifts' a low-score
user's post who later on becomes a moderator, then this
moderator should be rewarded. (because he predicted
correctly that a low-score user made a useful post)
The priority should not only be to filter 'good stuff' from
'noise', but also to reward people who are good
filters and dont mind to sift through filth to
find the gold.
64-bit is not necessarily a win. It has twice as
much icache footprint and takes up more memory
bandwith as well.