As an editor, I perform research locally and globally throughout the day and find many obstacles to the data I'm looking for. I thought I'd take some time to create a short list of those features a “semantic“ search engine should include:
1) Determine the word's part of speech (i.e. noun, verb, adjective, etc.) Oftentimes the same work can be used as a noun or verb (i.e. type, wind, singing, etc.)
2) Know the meaning of words that are spelled the exact same but mean two different things, even if they are the same part of speech (i.e. pool, mole, pry, yard)
3) Determine if the noun is the subject or object of the query
4) Not penalize the searcher for misspelled words. Based on the query, the search engine, for example, should know if I meant “aisle” instead of “isle” or “know“ instead of “no“.
5) Perform research - the other day (not at work) I wanted to know the number of Democrats versus the number of Republicans in U.S. Congress who have served in the U.S Military. I could not find this info already compiled. All I could find were each member's biography, which included military service and party affiliation. It'd be nice if a search feature could compile this info for me or at least organize the data so all I'd have to do is count. Basically, I'd like to create relationships through my search engine based on my queries, either locally or globally or both! Maybe I'd have political-party affiliation in a document on my local computer, but military service resided in various other places on the Web. The search engine could relate and organize data from everywhere into one result.
I know google does some stuff with misspellings, etc. but hopefully the others features (especially number 5) are not too far off.
If others would like to add to the Search Engine Wish List, either add to comments or post to your own blog and place link in comments.
Thanks!