Building a better Website and Logic...
I read this article on 19 things NOT to do when building a website. I pretty much agree - I mostly avoid websites that do many of those things (although I find pull down menus annoying, I don't think they should be outlawed entirely - if they work correctly - meaning they work just like tabs and the menus can be ignored ;) ). Do not resize or move my window, do not make me sit there thinking the slow loading home page is the home page only to discover it is some animated graphic you find amazing (and I find to be a yawn), don't try to impress me with your flash capabilities (you are presenting your site to a guy that has disabled animated gifs in the browser - I don't like pages that "move"), don't play music for me - I'm already playing music and mine is better than yours, if you don't work in Firefox - I'll use IE tab to see if there is a form to complain - but that is about it!
In short, a pretty good collection of "things not to do" - sort of like "worst practices". I do a worst practices talk about databases, it is much easier to do than a "best practices" - for the simple reason that worst practices are pretty much universally wrong - whereas a best practice applies just a little more than 50% of the time; in many or most cases - but not all. With best practices you have to do more work and caveat them - describing when not to use them as well as when to use them. Worst practices, so much easier!
I stumbled upon this fake motivational poster shortly after reading the 19 things article. The two seem to go together! Do many of those 19 things and you should just get that poster framed and hang it above your monitor.
Onto logic and false conclusions... I was reading this Dilbert blog entry which points to this NY Times article. Just because a set of experts in a field say something is true, you might still want to ask about the science. I thought the analogy to "Who wants to be a millionaire" with the "ask the audience" question was excellent. When the audience is polled at the same time for what they think to be true - they generally get the right answer. But, what if we polled the audience one by one asking for their opinion, what would happen then... It is highly likely that the slowly building consensus would take over and sway the audiences answer as each in turn spoke theirs out loud. Makes you question the "but everyone knows" statements - or at least it should :)
I'm still dumfounded at how many times people still say "separate indexes from tables into separate tablespaces for performance". Sigh, but everyone knows you should... So, there you go.