[pigsy]

I've run some test on the performance BEFORE and AFTER.

My test was conducted on an average page, and I recorded the "Page generated stats". I had 191 replacements, and using the DB cache.

The results were page loading time on average more than doubled, from an average of 0.5 secs, to 1.2 secs. This is an unacceptable increase, so despite really liking what this has to offer, I'm going to have pass on using it. If anyone can think of a way to make the performance better I'd love to take another look at this - unfortunately you can't get round the fact that for every new link added it has to check it against every word in the text, so in my case 1500 words x 191 links = 286,500 comparisons being made. Even on my dedicated dual-Xeon server it takes a big performance hit.

I wonder if a different approach can be taken, along the lines of how some sites you have words linked to ads - they don't seem to do anything with the html itself, but must be some clever javascript at work - perhaps this could be the answer.

All the same, praise to The Greek, for all his hard work, and if your only using this on a handful of links, I think there would be no problem in doing this - it's only when you want to have 100+ links you'll start to slow down signifcantly