Exploring search using keywords
By jasonslater • Apr 8th, 2008 • Category: TheoryTo understand more about the way search works, and to gain an idea into how it could work in the future, it may be useful to talk about what we would do if search was not available the compare it to a keyword based search engine. So, if the Internet did not exist - what would be the process for finding new information?
For example, if I wanted to know why black holes are appearing - what steps might I take?
Step 1: What
The first thing would be to decide what sort of information I need. For example:
- Do I have just a passing interest?
- Am I writing an academic paper?
- Do I need an image of a black hole?
- Is it for a competition?
- Have I seen a black hole and wanted to find out if it was dangerous?
- Do I have concerns about black holes in my immediate vicinity?
- Is my interest similar to black holes but not exactly black holes?
Step 2: Where
The next step would be to decide where to start looking for this information:
- Ask someone close to me
- Buy a magazine
- Contact a professional
- Telephone the local observatory
- Do a college course
- Call a builder
- Borrow a book
- Visit the library
- Watch television
Step 3: Why
After deciding what I am looking for and approaching one of the where locations - I will probably be faced with the question Why? This is typical for someone else to understand what my requirement is and for me to clarify my expecation, for example in the library:
ME: Can you point me to the black hole section please?
LIBRARIAN: Try the Cosmology section - why are you interested in black holes?
ME: Because I a have a problem at home with black holes infesting my walls.
LIBRARIAN: Oh, in that case you need the Biology Section - and if you are being infested then try household pests.
In terms of Search engines, the only words that I could use for definite at phase 1 for a keyword such are: “black holes”. Let’s try that in Google:

Ok, we get nearly 3.7 million results talking about black holes and, as expected, they have returned cosmological black holes as the librarian initially pointed out. However, the black holes are appearing in my imaginary house and I doubt very much that I have a concentrated hub of singularities.
However, by clarifying my expectation to the librarian I have obtained two new keywords “infesting” and “household pests”. So let’s try the keyword searches again with the new keywords:

This is more like it, only 16 thousand hits and the first article talks about termites infesting a house and leaving little black holes. Sound familiar? Reading the article a recommendation is to call a pest control company - a far cry from the phoning the local observatory as may have been an early suggestion.
Initially, this (albeit rather small example) tells us that keywords alone are not enough for finding informatino. Keywords may require two additional factors - context and clarification. In the example, context was used in the clarification process, e.g. “my house” and “infesting” which clarified the keywords and offered new keywords which refined the search.
Could a two stage search process work? If I type a keyword or range of keywords into a search engine would I be put off by a context clarification?
jasonslater is
Email this author | All posts by jasonslater


[...] Having been of the feeling that may be losing the war, at least we have won a few battles with incoming spam in our business. There are two main problem areas that we have been looking into. Firstly, there is the scenario of When spam is not spam which we have finally managed to bring under some level of control and reduce the nearly ten thousand daily spam email down to a handier thousand or so. Secondly, and trickier to deal with is When spam should not be spam. Key words seem to be one of the few options available for dealing with information and these have played an important in a recent experiment we did Exploring search using keywords. [...]
[...] The technology week in view: 14th April 2008 Having been of the feeling that may be losing the war, at least we have won a few battles with incoming spam in our business. There are two main problem areas that we have been looking into. Firstly, there is the scenario of When spam is not spam which we have finally managed to bring under some level of control and reduce the nearly ten thousand daily spam email down to a handier thousand or so. Secondly, and trickier to deal with is When spam should not be spam. Key words seem to be one of the few options available for dealing with information and these have played an important in a recent experiment we did Exploring search using keywords. [...]
[...] Exploring search using keywords [...]