Faroo: Peer-To-Peer Web Search

At first I thought, awesome, finally an open distributed search engine.

...Oh wait, you want me to install your closed source code on my computers, to give you all my personal data and CPU cycles for you to own and use as you like. (Which you "promise" won't be exploited or abused).

Make no mistake, this is a voluntary botnet for a single party to use to build a company on. This voluntary botnet will provide search engine services from a central point so they can keep the majority of profits without the overhead of paying for resources like every other for-profit company.

This is not Peer-2-Peer it is Peer-2-Company. Also they are late to the party on this approach, Microsoft already tried this with Bing: https://googleblog.blogspot.com/2011/02/microsofts-bing-uses...

This model is just begging for people to troll them by releasing tools to poison their database with nonsense data.

They should build a -real- open p2p distributed system (the blockchain is not perfect but a good example). Short of that if they want to be a for-profit centralized search engine, fine. They should go buy some servers and databases they can directly control. and try to be a closed source competitor to duckduckgo.

So would it be fair to say that effectively Faroo works by:

1. Scanning your browser cache for the sites you visit.

2. Setting a ranking to the sites in your browser cache, based on how frequently you visit the site.

3. Merging this site + ranking information with other Faroo users in a search index.

4. Distributing the search index in a distributed way, perhaps all nodes only having a fraction of the total index to prevent storage issues.

If that's a simplified version of what is happening then I could see it working. If I've misunderstood something let me know.

It appears they are relying on a combination of not documenting their ranking algorithm and updating it periodically to prevent spammers from flooding the system with bogus "attention" signals. I'll leave determining the likelihood of this approach being successful as an exercise for the reader.

I found this a little problematic:

FAROO indexes only pages which are located in the Internet, but no Intranet pages or HTTPS protected pages

If it simply can't see HTTPS pages, it'll leave a large chunk of the internet invisible to the search engine. I understand the reason for this, but it's a technical limitation they'll have to find a way past to make it useful as more and more sites encrypt by default.

Sorry but where can I try FAROO without downloading an app or connecting to an API?

Something similar was posted to HN a while ago[0]. However, it was made specifically for indexing scientific papers without keeping all the content. Totally open source and runs in the browser. I feel like a more generic implementation could be done well, but hasn't been so far.

[0]: http://juretriglav.si/an-open-distributed-search-engine-for-...

Sadly, by its very nature, it misses a lot of the "long tail" of rarely-visited sites that Google et. al. can crawl.

We tried using various for an all we developed and its results were fairly poor, so we had to switch to bing.

Gene Kan (https://en.wikipedia.org/wiki/Gene_Kan) also had the same idea, until he shot himself while working on a distrubted peer to peer real time search engine in front of his computer I think.

I heard about him when I watched a documentary about Napster.

No documentary about Gene Kan exists. He is virtually unknown but his work was an important contribution.