One thing I've noticed recently with sites that host fake files is that they are usually non-browsable.

Would it be possible to add an option to filter query results from such hosts and also to prevent downloading from them (the latter in the case where the fake has the same crc hash but is full of junk, hence corrupting your download).

To take it a step further, it'd be _really_ nice if there was an option to record all such IP addresses as they're encountered in a text file using the peerguardian/bluetack blocklist format.

You know don't you that these spam results offering fake or virus files are autogenerated by systems set up by groups similar to Overpeer. Google Overpeer to get an idea of what I mean. These groups do not use static ips. So you might ban specific or ranges of ip's one day, but they then switch to another range. Also those ip's you banned are then taken up by other poor innocent souls.

Try using your Junk filter as much as possible. It's a dynamic learning filter & takes time to be trained. For simple autogenerated results it's quick to learn to filter it out. But for viruses, it's a different story. You need to persist.

Yes it'd obviously be wrong to simply add the list of IPs to a blocklist permanently. I mentioned that more as a means of getting a list that could then be investigated (as some may be owned by a **AA company and may not be dynamic).

The concept of dynamically blocking within Frostwire both results and download/upload traffic from IPs that don't allow browsing - wouldn't that be effective without causing innocent users undue harm?

But you're suggesting browsing is a definitive method to achieve or research an issue. The problem is browsing is not a reliable tool. One minute I can't browse someone, 10 mins later I can. Perhaps initially with zero files, but then with 50 & later with over 1,000 files. Or other way around. One minute I browse them & they show 4 star connection quality, 15 mins later they show one or 2 star.

So if you're using browsing to determine who is a freeloader, that's a very poor & ineffective & inaccurate method. Browsing can be affected by how many upload & downld slots & bandwidth both you & them have, & whether perhaps they have chat enabled (which tends to make it easier to browse), etc. I might be affected by local traffic or traffic somewhere else between you & them.

I'm not sure what you mean by dynamic ip filtering, since their ip's can change every few days or so. So how to keep track of their new ip's, etc.

Well ... I've noticed that for me the "Browse host" option is always greyed out for fakers and I've hardly ever seen that for normal users. I'm not talking about being able to browse but getting a zero file return.

So what causes the greying out of the Browse Host feature for some hosts? Is it network latency, or is it something in their client that disables browsing? If the latter then I would still suggest that such hosts are not normal users, and are certainly not running LW/FW or Cabos. An option to filter search results (and to prevent downloads from and uploads to) such hosts can only protect the network and help reduce the overpeer problem imho.

Anyway, if its an unreliable method then that's ok, not the first or last time I've been wrong about something. Just thought it was worthwhile asking

Sorry to harp on about this issue but the more I think about it, the more I believe that an option to filter search results from such hosts would be a good thing.

I'm not talking about stopping leachers from accessing my files, but about a method to avoid downloading fakes. If I as a user want to be overly paranoid and restrictive about what I click on, surely I'd be doing no harm to anyone else, and surely an option to help me be paranoid isn't bad?

Along the same lines, it seems that many fakes spread via the "sharing incomplete files" option that most clients seem to have on by default. Therefore as well as not wanting to download or see files from hosts that don't allow browsing, what about only downloading from hosts that have the complete file?

LW/FW etc seem to work on the assumption that other users in the network are honest community members, however this is what the people behind overpeer etc use to infiltrate the network with piles of rubbish. I don't know how technically possible the above suggestions would be to implement, but in concept wouldn't they be helpful?

I believe the programmers are looking at the situation very seriously. See Sponsored Results (click on link) about one aspect of this. There's not a simple solution. It goes without saying it would be good to be able to overcome it more easily. It's been in the past 12 months these groups have been working extra hard at polluting & using newer techniques.