i have to agree that the current events category e.m.p artist certainly is a bit of an irritant. whilst i am certainly not a fan of trollophobes, sock-puppet conspiracy theorists, and xenophobic new-registrant tormentors & other bullies & crybabies and would-be censors on general principle, i think i see where this is heading and it is not good.
present company excepted of course, i would say "there are user names in the message headers, so skip them if u don't like what they are doing, and what they are doing is not 'spam' (unsolicited commercial e-mail) and leave other people alone and go do something else like watch moss grow on yer arse."
more respect for those with differing views and/or some more awareness of those coming from other frames of reference from themselves in about 45 per cent of i.d.p.f. users would also be helpful. i for one head first for the opposing viewpoint posts coz i already know what i think, and am secure enough not to need it re-inforced continually and shrink in fear from opposition, especially since they are often right on at least one point. if one is thin-skinned, insecure, and/or hot-tempered enough to prevent this, maybe they shouldn't be posting to and reading a current affairs board, eh?
in extreme cases, the originator of a given hijacked/progressively more obnoxious thread can go in and cancel it and repost it to get rid of the garbage.
the situation in question is a problem which seems to have originated at least as far back as the early days of the iraq war and leads to a situation not unlike a clogged drain because it is another species of tactical network abuse related to e-mail bombing and running bots on usenet groups and web-based boards (flooding) and is not an issue of content. remember otto/betty/shoeless/50 others and his/her counterbalance "god luvs u" from that time? it was a mess, especially since there were fewer categories at the outset so their/his/her output about whatever it was on that day gets mixed in with posts about the local spring non-partisan elections, the general election in germany &c.. it was a big mess. u had to slam a whole bottle of wild turkey, smoke a reefer as big around as a aerosol paint can, fire up that opium pipe with a piece the size of a tennis ball, or shoot up a big dose of morphine to sort through the board at that time.
sooo...the drāno in question would be require registrants to look at their e-mail for a message from i.d.p.f. which contains a link to an account-activation page, if this is not already being done.
i would propose having the code for the programme* which runs the activation page look for similar headers such as mailer, user-agent, path and others and then parse the e-mail address and generate a numeric score for level of similarity, and do the same with http headers then have it sideline for manual processing:
1. anything scoring high for duplicate e-mail headers within a given time frame, probably using either an arbitrary number or a figure derived on the fly by the programme based on standard deviation or something.
2. registration attempts from known and reasonably suspected problem domains, especially if it is a web-based free e-mail service (u have to publish this list somewhere in i.d.p.f. for the benefit of potential innocent bystanders and others)
3. non-existent domain names -- e.g. those which return an error when the software attempts to do a brief traceroute (one hop should be sufficient)
4. a certain number of posts in a given time from the same source, perhaps counting lines and adding up the character code values to get a line count and checksum and then sideline for manual inspection blasts of similar posts and/or those which came in at above a certain volume in n number of minutes.
5. actual duplicate e-mail addresses, of course.
6. those above a certain pearson product-moment correlation coefficient comparing the time and number of posts per day for two or more suspected "sock puppets"
on the other hand, anything involving publishing i.p. addresses from which the http get and puts originate would seem to be effective at first knee-jerk, but it is not, primarily because this information becomes intrinsically less useful for such a purpose every year, even month, it seems, for several reasons.
a big issue for i.d.p.f. would be that doing so would create a bifurcation betwixt technology haves and have-nots. not only has the technique of using proxy servers, shell accounts, scripts to screw with the packet headers, and other tools become proletarianised to the extent that large percentage of users of any given site cannot, in fact, be accurately located by an i.p. address and its associated machine and domain name. attempts to do so, if they do not immediately result in a dead end, often wind up including innocent bystanders such as i.s.p. staff or even people who those on the hunt incorrectly think it was. a lot of firewall programmes also remove or scramble http request headers as well.
and i would hope that the powers that be know better than to publish actual e-mail addresses since those who want to get mail can already choose this option in the setup screen and the tone of some of these threads makes it easy to imagine net.terrorists going after people who look at them the wrong way...well, the cyber equivalent thereof.
* there is more than one way to do it® -- but i propose using perl for the main body of the programme, or perhaps using a unix shell script as a wrapper and go-betwixt with the html and php of actual web page itself, that does all the string, numeric, and network operations, and if the virus scanner on this server does not summarily zap code in posts, i could very well post the code for the main algorithm and/or whole programme we use to prevent this kind of thing as well as some alternatives such as vbscript or rexx.