The Chronicles of Stuffed Guys

We’ve finally released the long-awaited version of Stuffed Tracker with reports caching.

Ivan actually started to work on caching back in February or January 2006 (don’t remember exactly now). So it took a really long time to finish as you can see (though Ivan has been constantly distracted by other internal projects).

I really like the end result. We’ve managed to combine the idea of static reports which are updated continuously on the side of the server with our fancy dynamic reports constructor.

To be honest, it wasn’t clear for us how to pull this off with the reports constructor still being present in the product (and it is the main way to work with the aggregated statistics in Stuffed Tracker). But after a 2-month testing period of the new version on our client’s server with a tracking database close to 1Gb in size, I can now safely say that the new caching feature works great and it does solve the “slow reports” problem that any statistical software is plagued with by definition.

Working with the tracker that has to make queries in a 1Gb MySQL database is now fast and convenient (with caching enabled of course).

We’ve even introduced a so called “smart caching” feature. With this feature enabled Stuffed Tracker constantly analyzes how long it took to prepare each report and displays fast reports in real time and slow reports through the caching feature. This effectively means that you can enable smart caching and as soon as some of your reports will become too slow to generate they will be cached, while the others, which are still fast, will still display in real time.

With caching problem gone, we now have only two major problems that have to be solved:

1. Complete deleting of the old data (probably, this will include automatic deleting via cron script) — to solve the problem that the database grows too big and doesn’t fit into the accounts of some of our clients.
2. Archiving of the old data — this one should obviously come together with complete deleting, because it’s not good to just lose the data completely, even if it’s old.

These 2 problems are much easier to solve then caching and I am thinking that at the very beginning of the next year we will get rid of them.

This all means that we are finally reaching the stage in the Stuffed Tracker development when we will start to add tons of new exciting features instead of just solving the current problems.