This would answer the question "why the government does not want us to have this". But my question was completely different.

We have plenty of capable programmers who don't mind "illegal" data sharing, and we have millions of people who disagree strongly with the current copyright laws, for example, and having no realistic means to change the law, choose to ignore it.

Let's set aside and not discuss the moral and legal side of this right now. We know many people exist who would support and enjoy such a system. So why has no-one managed to build it so far?

P2P poisoning is done by malicious servers that announce millions of connected users and return results for all sorts of keyword queries with fake filenames and spammy or malicious content.

I was suggesting a reputation system based on signatures as a possible part of a solution to this problem in P2P networks. The rating system was not the main goal of my suggestion but a means to end, so saying, "But some torrent sites already have user-rating systems," is missing the point.

What about a pseudonymous reputation system?

Let the uploaders voluntarily sign their uploads. Then, as the user verifies a few packages by that uploader, they assign a higher trust score to them, so their personal spam filter ranks this uploader's other packages better.

I can also have friends and trusted vendors whom I can whitelist based on their signatures.

We can also have global distributed "trust rank" for uploaders and individual packages, based on feedback from multiple users. That's not very reliable, but possibly useful as one of spam filter heuristics.

Then, we can also have pseudonymous meta-moderation, where users who consistently flag good and bad packages are trusted more about their future feedback.

So maybe this community screening/flagging could be decentralized?

And maybe we could still find a meaningful set of heuristics for the spam filter? Signals from users, file type, file size, first uploader location and software stack, known file checksums in the package, etc. - maybe if there are 100 carefully calibrated input signals, the detection rate would keep it usable.

Also, AC in this thread made vague suggestions about using PGP authentication (for trusted uploaders?) and some bitcoin-like algorithm (which I unfortunately don't understand how it applies here, but maybe it makes sense to someone who creates this sort of systems).

I obviously understand very little about the technical details of the problem, so I'm just trying to ask questions and pass around the ideas.

Another useful idea, probably from the same AC as above:

Now this is ironic in an AC post. Also, I don't understand your point about anonymity working against you in preventing the monitoring. Are you talking about "circle of friends" type networks?

I also noticed the spam/"poisoning" problem with p2p networks. Using BitCoin-like algorithmic approach to solve that is a new idea to me, which I need to think more about to understand it, but it does sound interesting.

Another decentralized system where spam problem was somewhat mitigated is email. For me, SpamAssassin with its bayesian filters was a 90% fine solution, and GMail with its power of scale solved the problem for 99.9% of cases. There may be 1 in 1000 slipped spam or a false positive, but that doesn't bother me much, and it is uncomparable to the spam disaster we had in ca. 2000.

So spam and poisoning is a real problem, but not an unsolvable one.

This sounds like one of plausible reasons. I imagine this would keep me from joining one of those networks where everybody must store and share a bunch of encrypted file fragments without knowing what's inside. I would not like to facilitate in distribution of certain content in any way, nor to be potentially liable.

But is this so fundamental? Perhaps we could develop some system where everyone can self-moderate what they share. Maybe also some sort of voting and commenting system could help.

Building such a system is easier now that we have bittorrent and DHT. A keyword-based search layer on top of that is a smaller problem. Then, if someone is concerned about privacy flaws in bittorrent, they can also try to use it over some encrypted tunnel.

There is so much wrong with this post:

1. File sharing in general is legal. Sharing certain specific files may be illegal in certain jurisdictions.

2. The legal side of the problem is separate. In many spheres of life there is an area of untracked relations between people. Small cash transfers, personal presents, favors, discussions, meetings. I understand that the governments want to control, censor and tax all of this as much as possible, and at least ask for voluntary self-reporting in many cases. But I feel this "breath room" is important to keep the society sane, and we should have an equivalent in the digital world.

3. The software market would not tank even if file sharing became easier. There are risks in downloading software from untrusted sources, and people who can afford it (or cannot afford the risk) normally buy it. Then there is also support, upgrades, special deals and so on. Media market has more to worry about, but also not as much as they claim.

Yes, I know. I was referring to the bittorrent community in general, including the index websites, forums with links, seeders, etc.

In fact, it makes me wonder even more. When we already have a great data transfer protocol, and even a distributed database for hashes that makes "magnet links" work, why is there still no mainstream distributed keyword search system on top of that?

Part of the reason could be people disappointed with spammy eDonkey experience. Another part, as one poster here mentioned, is that people don't want to join networks which may be distributing objectionable content. But both these reasons don't sound so absolutely unbeatable by some creative filtering and public moderation.

Granted, this whack-a-mole game with individual torrent sites makes for a fun show sometimes. But I find it embarrassing that the online community has to work around these issues time after time, and that some good people get caught up in legal battles.

Are there any good alternatives to bittorrent for private, anonymous file search and exchange? I heard about several "darknet" projects, but they never seem to gain traction for some reason. Given a huge number of hobbyist hackers who support free exchange of information, I am surprised.

Is there a fundamental reason why we cannot have free, anonymous file exchange? Or is everyone just happy with the status quo?

I would put the telecom second and the media distribution mafia first.

It is pathetic, true, how the telecom providers have been selling a commodity service on mass scale for 20+ years, yet the pricing and service quality are on "novelty" levels or worse. Your cable bill has no good reason to be higher than that (local) phone bill 30 years ago. One of the reasons for the pathetic prices are the unreasonably high media licensing fees and unbreakable channel bundles. The cable companies then cut costs on everything else, which gives you multiple week waiting times to connect, half-hour wait times on support lines, and clueless staff.

And the media mafia also criminalizes everyone for downloading a few songs on P2P and threatens with lawsuits.

I did pay attention. It is one thing to get net positive energy in an experiment, and another thing to capture that energy and to sustain the reaction in a feasible way.

I would argue in favor of this experiment for the possible interesting scientific results, but by trying to market it as a viable power plant before 2020 they are turning it into a scam.

Not if they really had made their case.

Are you going to invoke the oil industry conspiracy, too?