Ah, thanks. I never considered JS - interesting idea. Looking at their code in the example I gave, the counts appear to be prepopulated in the original wget source so in this cause I guess it's server side.
I'm just trying to figure out how this can be done fast with 200k products, each with 50 odd attributes in a joined table with a regularly updated product record set.
I guess the third option would be to perform one mysqlquery, get all attributes for every current product then increment variable variables in the while loop and put the load onto PHP rather than MySQL.
I know there are various ways to achieve this, but if your target was speed without scaling out the hardware to read slaves during peaks how would you guys do it in the example I gave please?