Testing; where's the problem?

So, we went live last Tuesday morning, with a sparkly new tutorial suggested in part by prof and gang. We didn't fully publicise it, but released it to a select group of testers.We edited the interface a little based on their comments, and modified /added a few features~

And then... we hit an interesting issue. Users (and one designer) started complaining of lag that ranged from mildly acceptable to extremely long. While neither Eldwin nor I are really experienced in optimization, we didn't think our code would scale *that* horribly, especially when there are, as of date, 24 users, 3 of which seem not to have explored anything beyond our tutorial, and 2 who we can't even make that claim.

So, the logical explanation might be... geography. Most (all?) of our testers are based locally, and AWS is... not. And considering how often our app makes http requests to our AWS instance, it's not surprising that some manner of lag seeps in...

Last Monday, a day before we went live, we had a presentation by SoC's own Lai Zit Seng. Zit Seng is a very skilled... probably to pigeonhole him as a network administrator might be an understatement. I met him before during a training session for network helpers held in SoC, but this time I was really really impressed. Listening to him, optimization is more of an art than a science. You can't fully learn it from texts or lessons. You will have to try it, play with it, having it blow in your face (with backups, hopefully) for it to work and you to learn. He has a lot of experience in his field, and I am very impressed with it. His presentation slides are on my to-read list half a year from now, when our app takes off and we actually have hundreds of users (at least) playing it =P Like prof said, to have a problem in optimization is a good thing ^^ So, back to creating more features that the users (and 2 crazy designers) have requested for~