All,

Not sure if I'm posting in the right forum, or if there's an official 'wish list' I need to add to, or if this has been raised before. If so my apologies and please advise where I should post....

I have an idea for improving the search function. At the moment the code for determining what defines a searchable 'word' is a simple split statement that uses a space character as a terminator. So a 'word' is a block of charaters terminated by a space. But by my reading (and testing) the word 'space' at the end of the last sentence is not the same as the 'space' in the middle of the previous one as one is terminated by a fullstop and the other not. Clearly in any natural language they are the same, and should both go into the SearchWord table as different instances of the same word. Currently they don't.

An improvement would be to make the definition of a word to be a sequence of characters terminated by one of a number of characters. I would propose that as well as the existing space  you would include .  , ? ! ; : ' " and )  - even neater would be if you could define these in a steering file that could be modified if required on a site by site basis to allow for other special circumstances.

Just a thought. For the moment I've done a quick and dirty mod to the site I have hit a problem with to work around the issue. What does anyone else think about it?

Regards,

Graham