python with pytables for the etl, perl for the parsing and page caching, and php to display the results I would like to make something with erlang that could be distributed onto home pcs and create a giant botnet search engine that would have 9 9s uptime, but I've been working 60-70 hours a week at my day job, ughh...
Does it have a unique web crawler? Looks like you already have quite a few pages cataloged. -- Nathan Enns
It gets its results based on the stories submitted to http://vi.goro.us/ (its a different subdomain) and crawls through the links of submitted blogs every couple of days (around 800-1000 new/day). It doesn't refresh its results nearly as often as the big search engines, but it works good enough for what I search for (coding and corvettes!). You can tell its really biases towards sites with a lot of news articles since they are submitted the most.
It's got all my sites indexed, even the ones with 0 backlinks that I've submitted to only the major search engines... Are you sure this isn't just another front end that uses results from a real search engine?
Its not completly unique. If it knows about a site and has given it a score (it bases scores on user input, google trends, yahoo buzz, slashdot, and a few other sites) then it ranks according to that. If it has never seen a site that uses that keyword than it searches several search engines adding specific syntaxes to find the most appropriate results and displays. After that site is saved off it gets assigned a score within 2 hours (a job runs that does the comparisons every 2 hours). The whole thing only took me a couple of hours to build and was more for fun then to be a main search engine.
it's simple but the url is complicated not too many people will remember it. And the search results are limited. My site is also not indexed lol.. Anyways I like the clean look
I have checked the search engine and it has simple design and layout. In terms of result, I got a great result.
I haven't decided about sponsored links yet. Maybe down the road. How does it work? : It uses articles from vi.goro.us first (user submitted) and if it doesn't find your search term it looks at some major search engines. If it doesn't find it with my specific criteria than it lets you know. The articles are then ranked by a score that I determine with google trends, yahoo buzz, slashdot, etc.. to determine there placement on the page. I just added a current search underneath of the input box that lets you see what people are searching for.
Interesting. Results are different from the ones showed by Google and Yahoo. I also like the fact that every page displays a lot of results - although Google and Yahoo can be set up to do that... - Just a question: How can a user submit a page for your engine to find it?
For those of you with php sites, try out the new API. I tried to make it as easy as possible to get started making your own search based on mine.
If you are willing to pursue this, I might be willing to help. I love Erlang But to make life (a lot) easier, just install; http://lucene.apache.org/hadoop/ ; it works like a charm; I am using it for everything these days. Brilliant work of those programmers.