DIY search engine takes on Google

[Turborama]

Source: BBC

November 29 2011 Last updated at 09:56 GMT

Google, Bing, Yahoo and other search sites have a new rival called YaCy.

Backed by free software activists, YaCy aims to literally put search into the hands of users by distributing its indexing engine around the net.

Anyone can download the YaCy software and help the search system improve and spread the load of queries.

Its creators also hope YaCy will be much harder to censor than existing systems that pipe queries through centralised servers.

Read more: http://www.bbc.co.uk/news/technology-15935550

---

The article goes on to state that the YaCy search page was opened to the public on 28 November. I can't seem to load it the link they give to to it, though: http://search.yacy.net/ I use Chrome and all I get is "HTTP Error 500. An unexpected condition was encountered".

[Fumesucker]

1. Chrome didn't work for me, Opera did.. n/t

[RKP5637]

2. Firefox gave me a blank page initially, but after several tries worked. Opera works OK.

Message on YaCy server http://search.yacy.net is:

"There is currently massive load on this page because of press releases today. A search might be not successful now.

The search results will be provided by distributed peers from the 'freeworld' search network (which consist of private computers).

This is just a demo, not the search portal.

We like you to use your own peer to search within the YaCy network. Your search portal is on your peer.

If you want the 'real' YaCy, install a peer on your computer and use your own search peer locally.

Because there is so much load on this page we need more people who provide their peer as portal.

It works like this: there are only a few peers which provide search 'as face to the network' to this portal, but if we want to scale we need more: YOU can be one of that search providers.

Just post your peer address:port (with a static IP or host name!) into our forum topic about this portal and we will add your peer to this search portal!

YOU will then become the search engine for EVERYONE ELSE through this portal!"

[MannyGoldstein]

3. Chrome worked for me, but...

I get a message that says "There is currently massive load on this page because of press releases today. A search might be not successful now."

[Felinity]

4. Good news

Now would someone start a new YouTube, since google has ruined it.

[onehandle]

5. Fantastic. The big three, especially Google, need to die. Google is trying to 'own' the Internet.

Google is the biggest enemy of anonymity.

[saras]

6. Not the worst, just the biggest. Imagine how much worse if AOL were on top.

[PassingFair]

8. Or AT&T...

:scared:

[1monster]

7. I got it... It has a message that it is over loaded due to the press release:

There is currently massive load on this page because of press releases today. A search might be not successful now.
The search results will be provided by distributed peers from the 'freeworld' search network (which consist of private computers).

This is just a demo, not the search portal.

[Name removed]

9. Deleted message

Message removed by moderator. Click here to review the message board rules.

[octothorpe]

10. Sounds interesting. LOL @ the deleted message... I'm curious what could have been said.

God damn rabid googlers.

[rosesaylavee]

11. tried it today... not impressed and it was very very slow.

[octothorpe]

12. Same results... Maybe in time it will become something useful

[rosesaylavee]

14. Could be... I'll try it again when it goes beyond 1.0

[Xithras]

13. YaCY = Neat idea. Terrible implementation.

Because the search algorithms are only customizable by programmers, most people will be left with a 1990's level result-set. It will be like using Lycos again...10,000 results for every query, and the three relevant ones will be buried on page 42.

On top of that, it requires a client side installation, wants to consume 30GB of your disk space, a permanent percentage of your Internet connection, and many of your processor cycles. Performing searches also spawns a complete webserver instance, introducing potential security holes, and unavoidable performance problems. Not only are many going to be unwilling to do that, but many of those who ARE willing will be unable to do so because of hardware limitations (Running a tablet? Mobile device? Thin client? Laptop? Sorry, you're SOL).

The biggest problem with YaCy, though, is that it's core spidering concept is flawed. YaCy provides users with two different indexing options. Crawling mode will burn a huge amount of processor time constantly, making it unsuitable for most computers. YaCy's "fix" for this is to offer Proxy Mode, which simply indexes pages as you surf. In theory, as millions of people surf the Internet, the whole thing will gradually get crawled. The problem with this model is that it naturally favors popular sites, while placing less frequently visited sites at a serious ranking disadvantage. CNN.Com will always be up to date, but the blog dealing with local environmental issues in some random midwestern city is going to be lost in the noise, because its comparatively low traffic levels won't allow it to be indexed with any real frequency.

It's a cool academic project, but it's not going to replace Google anytime soon.