On Thu, 5 Oct 2023 16:52:03 -0000 (UTC)
Post by Kaz KylhekuPost by Ben BacarissePost by Anton ShepelevPost by The DoctorWEll now more calls at East-asian spmatrollers are
flooding comp.lang.c to no end!
Is it a coincidence that it stopped here in CLC several
hours after I logged in to Google Gropus and reported some
fourty Thai articles as SPAM?
Probably. Ray (who runs Eternal September) is aware of the issue
and is continually (or so I imagine) updating his filters.
I have another feed and, using that alternate view, comp.lang.c has
had 2407 posts in the last day.
We need a whitelisting solution, at the federation level.
Yes, and no. Yes, whitelisting could be one option. But we already have
a defacto whitelist by the peers we choose to synchronize with. When you
add a peer to your config you have just whitelisted that peer. Managing
users in a whitelist fashion would be cumbersome and error-prone and a
vector for administrative abuse.
Whitelisting via the path header to exclude messages with unlisted peers
in the path would be viable. But you would need to sync with a dozen or
more peers to get all the good messages from a pool of much more than a
dozen peers. I think some sysops don't want to maintain a dozen or more
peers.
Post by Kaz KylhekuThat is to say, servers like Eternal September should reject all
articles from Google Groups, except from users who are whitelisted.
Unless complex cryptographic signature verifications are put in place
in the proposed central registry, forging would be a route around
this kind of whitelisting.
Post by Kaz KylhekuUsenet could get together and provide a registration service whereby
Google Groups users can register their account for whitelisting.
Those who don't are effectively shadow banned by default, confined
to the Google Groups echo chamber.
I reject the idea of a central registration service. That would be
severely abused to track and censor people. It might not happen right
away, but eventually it would morph into the norm.
If a peer is not making good faith effort to thwart abuse then every
user on that peer should be blocked. You should never try to be mother
hen over the users of someone else's service. If they are not smart
enough to detect the danger they are eliminated from the chicken
species. If mother hen is a failure, her chicks get eliminated from the
chicken gene pool, which is a brutal necessity of nature. Surely it is
not fair. So what?
Post by Kaz KylhekuObviously, that registration system would have to be better run than
Google's own account registration, which is the point. Google sets
such a low bar that effort would be needed to bend down that far to do
an equally bad job.
You are suggesting a central network registry, which will eventually
get taken over by bureaucrats or buffoons to everyone's detriment. It
won't "might" happen. It will happen. And then you also have a single
point of attack for government agencies to silence dissent.
Which country should the registry be based in? See the problem?
If a independent registration platform were used, in which peers could
choose which registry to use, that would be almost viable, as it would
give sysops the choice of association. But still the organs of state
could micromanage it for easy censorship.
Anything that is not sufficiently decentralized will be attacked by the
organs of the state or the organs of crime. Google is facilitating both
attacks. By treating Google as a sacred cow, sysops are effectively NOT
operating a decentralized protocol. Relationships change. Character
changes. You adapt to those changes. NNTP facilitates that ability to
roll with the punches.
Post by Kaz KylhekuOne technical difficulty is that, in the article headers, Google
Groups accounts are identified by an anonymized account ID string.
Users don't know what that string is, and there might not be any way
to get it other than sending an article to another server and getting
the headers from there. (Perhaps GG lets you view the raw article,
with headers, if it's your own).
You can block Google. All non-commercial peer operators should have
no problem with this. Google is the biggest source of spam and abuse,
and has done nothing for many years to fix the problem. You gain
nothing by allowing Google to continue abusing the network. You gain
everything by a concerted front to tell Google to take a flying leap
off a ledge. By spamming you all for years, Google has been telling you
the same thing. Are you going to take that any longer?
Look at it this way: Google does NOT care about you or your server or
your users experience. Why would you want to play mother hen and show
care and concern for Google's users? That's like a father ignoring his
own kids and letting them run riot with a cannibal gang, and taking the
neighbor's kids to the amusement park. Sure, the neighbor kids think
that's fair.
As for whitelising, it can only work non-abusively with a small network
of peers that are able to quickly agree to adapt to attacks on the
whitelist registry. Otherwise a whitelist registry it is just a
foundation for a new set of problems.
--
__|__|__|__ ***@sugar.bug
__|__|__|__ https://sybershock.com | https://syfershock.com
__|__|__|__ news://alt.sources.crypto
| | | #CipherTag #WaffleTag #Cryptologer
