This is a project that I've been working on and off (mostly off) for some years now.
The idea is to create a local search engine for just this site and to actually do the search locally on your browser.
Now I can't claim the original idea was mine, I first saw the idea in action in the o'reilly book Javascript application cookbook (now out of print). The search engine they presented here was pretty limited, mainly by the limitations of the javascript language of the time, but it did inspire me to have a play with it. I used the script from the book pretty much unchanged on my first bit of commercial freelance web design work I did. Well I say design, I actually reengineered a site for a now defunct system builder, called Compuadd. The site had been build in front page and was worked fine in IE, but didn't really work so well in anything else (these where the bad old days of the late 90's, I think).
Anyway during the reengineering I added the javascript page search, which given it's limitations worked fine. I later used the same code on a couple of other sites I worked on, each time tweaking the code to add more features. By the time I was coming to add the code to one of my sites, it was becoming clear that I would have to do major work on the code to get the features I wanted, namely the ability to limit the search to sub sites and some sort of google like ranking of the results.
I got bogged down in this rewriting, the site never actually got put up and the domain expired, so the work got abandoned.
Fast forward a few years and I started to hear about the cool new technology called Ajax, javascript and xml. As I started to read about it I could see that this would give me a way to have sub site searches, although I don't actually need those now, and generally make the search work much better.
I started to play with the basics of fetching xml using javascript, but then I discovered the many wonderful javascript libraries that where springing up all over the web. Libraries like YUI (Yahoo User Interface) and Dojo. These great libraries took a lot of the 'heavy lifting' out of the way, so I could focus on the search code it' self.
At this point I guess about 60% of the code is there and in my test environment it's working fine in a limited way. The main problem is building the search database from all of my web pages. At the moment I have no way to do this automatically and I'm not sure that I want to automate this, since I think it needs the human touch. This dose mean I have over 250 blog entires to index, this is going to take a while!
However the project is getting close to completion now and as soon as it dose you will be able to find the source code here as well details on the hows and whys of the design decisions I've made.
So watch this space!