Andy’s articles are one of the reasons I like blogs so much. These days most of my learning comes from blogs due to their easily readable, small bite, immediately available format. I find the published formats less attractive these days (sadly I usually have a pile of FoxTalks, Advisors and other mags gathering dust on my desk). But that is a discussion for another day…
Back to Andy’s posts. The big point for me in Part 1 is about knowing your user base. Are they pretty smart? – yes, then don’t treat them like idiots – I totally agree. I have a slight addendum that I’d like to make though – if you do have idiot users (and we certainly do for some of our products) then perhaps make things like the Delete message box an option (most products have a [ ] Confirm on Delete option tucked away in their Options settings these days).
However, Part 2 is where the real interest is. I am totally stunned with Andy’s findings on the effects of controlling indexes. I’ll be hunting through code tonight… imagine making a few tweaks to your setup to looping code and smashing the run times. I’d be putting that on the feature list for the next release… ‘The Sales Analysis report is now 45% faster’ – that’s got to be worth the upgrade price {g}.
Back in Part 1 Andy discusses the trap of over-informing – yes, we’ve all written that screen progress update code that is both (i) completely unreadable and (ii) slows performance. What intrigued me was how big the performance hit was. 15% is pretty substantial. And I’d be guessing that most of us developers have machines with decent graphics cards that absorb some of the overhead. On the ‘normal’ user’s desktop these performance hits are probably greater. More code hunting tonight…
Check out the articles here:
Andy Kramek: Writing Better Code (Part 1)
Andy Kramek: Writing Better Code (Part 2)
