Comments on: Catching up

By: dan

dan — Thu, 28 Jun 2007 21:25:46 +0000

nice, i’m glad you like it.

i was thinking a couple of things that seem apparent but might have made the list, more of an outsider’s point of view:

High Expectations

you’re only usually going to find a solution somewhere between half-arsed and the upper limit of your expectations – this gives you a lower distribution for final outcome vs. very high expectations of what is possible. (based on they thought they might be able to write tetris in one line)

Challenge All Assumptions

i get the feeling they didn’t hold back challenging all aspects – maybe they took this for granted. from my experience, often the biggest breakthoughs are based on this.

cheers, dan.

By: Dan North

Dan North — Thu, 28 Jun 2007 19:00:34 +0000

I only skimmed it but I already like it a lot. I’ll have to look at it in more detail. I particularly like the contrast between “marking” and “hammering”. They are great terms for something that seems to happen a lot with performance tuning. I also like the idea that he is flexible with his metrics. He measures things and then uses the measurements that seem to yield the best subjective improvements. Great stuff!

By: dan

dan — Thu, 28 Jun 2007 10:14:49 +0000

hi Dan,

re: performance tuning techniques write-up

you may find this useful – it’s a write-up of techniques for writing very small programs (disclaimer: by my brother); i’d always thought of it as a curiosity, but it just struck me there are a lot of similiarities with performance tuning:

http://www.survex.com/~olly/dsm_rheolism/method.html

cheers, dan.