On Sat, Mar 17, 2007 at 10:18:34AM +0100, Jerome Flesch wrote: > > Interesting, but most of the time moderation on the level of posters is > > a lot more interesting than on the level of messages. What might work > > well would be a true web of trust on Frost, where you can set it to only > > show messages a) from people marked GOOD, or b) from people marked GOOD > > by people you've marked GOOD (etc). > > > Firstly, that's not what I intended to design: > In the end, what I would want is to be able to moderate quickly a > specific message. Why ? Let's take Aku for example : This guy makes > often stupid trolls about Linux / Windows ... but sometimes he (or > she?) says some interesting stuff (sometimes ... :). That's why I > don't mark him as bad. But honestly, if I could get rid of his trolls, > I would do it. > Moreover, with a web of trust, there are two possibilities : You can > accept to read messages from new-comers, then what forbid someone to > create a new identity each time he/she send a message to be sure that > most of the users will get his/her message ? Or you can refuse to read > their messages, then it finally will make difficult for new-comers to > be marked as GOOD by someone.
Not everyone would have to read messages from newbies - just a few people. And yes beating spam is hard. But you don't solve this problem with your architecture either. > > Secondly, some people on Frost already asked to bback if he could do a > web of trust like that (but I don't know if he plans to do it). Ok cool. > > Another idea suggested on the Frost board is a retro-moderation > (something like: people creates a message black list, and other people > can decide to trust this black list). But it won't give any protection > against message flooding (as the web of trust in fact). A web of trust would protect against flooding. Well, sort of. It would protect against it on the level of avoiding spam being displayed, but because of the architecture Frost would still be vulnerable to DoS by inserting lots of invalid messages. This can be solved with web of trust by having each trusted identity publish both its trusted identities list and pointers to recent valid messages? -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Digital signature URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20070317/fc2273ae/attachment.pgp>
