Topic awaiting preservation: search engine (Page 1 of 1)

i seen a message posted asking how to create a search engine from scratch. i just need a bit of advice about this topic. As part of my final college year I must complete a significant computer development project that addresses a specific business or organisational need. I have 6 months to do this, and i was toying with the idea of a search engine. my skills include java, visual basic, access databases, and some sql. i have no knowledge of php or asp. do u think this is a viable project for me. also i would have to throw a slant on the project so that it is unique? or does someone have another significant computer development project that addresses a specific business or organisational need just off the top of your head, u never know!

bones

I don't have much experience with Java, but if its a robust language then you should have no problem. Most of your work will be done defining your algorithm since there's alot you have to consider. It should be relatively easy, although a lot of work is needed.

As far as a unique slant, you'll have a hard time coming up with something satisfactory since I doubt there's much left that hasn't been done by google or other search engine companies. I suppose you could choose to focus on a particular area where little work has been done or with specialized needs such as a phonebook search engine with literally millions of indexes or even a smart name search that can recognize common abbreviations such as Bob for Robert or Dick for Richard.

well, looking at what you claim to know... I'd probal ysay that this isn't appropriate for you.
Java is a nice language for buissness modelling and a lot of other stuff, but when it get's down to quick & dirty strings processing, there are quite a number of better alternatives.

You'd need to get up to speed on the sql to... after all you got to store your found information somewhere.

the 'name search with alternate spellings' could be interesting... you could even try to get it to differenciate between people with the same name. but beware, you'd easily do work for a doctoral thesis if you managed that.

bones: Interesting topic - one which I've been vaguely mulling over for a while. See, for example:
http://development.gurusnetwork.com/discussion/thread/2199/

I presume you will need a robot/spider to do your indexing and you might pick up some insight from looking at freely available site search engines which do this. Even if you don't use the specific language to make your search engine it might provide some insight.

Have a look at:
http://phpdig.toiletoine.net/

Also the FOAF people have developed a few robots to find and process FOAF files see:
http://xml.mfd-consult.dk/foaf/explorer/

Pos that might be an idea for the unique spin - with the increasingly Semantic Web there needs to be various tools to deal with this so that might be one idea..........

[edit: And there are always niches that can be exploited - just look at technorati:
www.technorati.com

which indexes blogs which link to other blogs and can tell you how many link to yours]

___________________
Emps

FAQs: Emperor