No doubt keywords and tags play an important part in the modern web. But what is a keyword and what is  a tag and what is the difference between them. My explanation would be this:

• A keyword is a common word used within a web page
• A tag describes a conceptual idea about a web page

As an example, take this paragraph from my website at www.jasonslater.co.uk

These days I have Google Analytics installed so any time I want to check on the relative success of a post or topic I can refer to that so what is Alexa doing for me? I also use Technorati from time to time as another useful measure though I tend not to check these sites as regularly as I used to - as most of my time these days is taken up researching or writing the posts (Blog overhaul update … it’s all about communication!).

Keywords may be: post, technorati, alexa

Tags may be: ranking, analysis, readership

The reason for the distinction here is that a search engine can easily extract keywords from a web page, and build up a word library for searching but surely it cannot, without some help, automatically create and assign a tag to a web page. With this in mind it poses the question what are meta-keywords in html headers?

Keywords are fixed, finite and relate directly to a post and so are straightforward for building into a keyword search engine however tags are not so easy to handle because the user conducting a search may have a different conceptual idea in mind for the same word. Take the example above, the tag Analysis after reading the paragraph is probably taken to mean traffic analysis or readership analysis but if I type analysis into a tag based search engine it can’t be sure what I’m looking for -> psycho analysis, systems analysis, market analysis? If the search engine returns to wrong results enough times the user may turn to a different mechanism for finding information.

EggnChips therefore needs to start out with a keyword analysis routine (the extraction of useful words from web pages) in order to build a library of keywords which can then be cross-referenced with a list of web page addresses (URLs).

This suggests that EggnChips needs URLs to work with. The idea of creating a bot to go out and fetch pages is intriguing but probably a little too far down the road. Initially we should probably focus on user submitted URLs and work from those. Mental note -> what are the ‘rules’ for operating a bot that automatically retrieves web pages?