Yesterday, Ed helped me word a "law" that I use for important decisions and that I see show up quite a bit in a number of places. It's the law of human relevancy.
The Law of Relevancy
No matter how relevant the information is, it's more relevant with the help of the right people.
The Human Shepherd
All this law really means is that no matter how well you organize and display information, at some point, there's a glass ceiling on how much easier you can make it for somebody to find what they need. There's always a place for the human shepherd.
Usage / Examples
- The obvious example is the Web 2.0 movement -- where people are the shepherds of the read/write Web. There's lots of needles in the world wide haystack and I'm glad there's people, voices, blogs there to help.
- I like the pattern on the Web where sites have a live chat to use a human to help you match their info to your needs, in real time.
- I like how Second Life provides the ability to "invoke a human" over just self-help and forums. (I proposed some models for "human help" in Visual Studio and as a general platform some time back I need to revisit)
- There's lots of implications for an Enterprise 2.0 world, but I'll save that for another day.
In an early version of Practices Checker (a tool meant to verify your solutions against patterns & practices recommendations), we tried to figure out relevant guidance based on the type of project (Web, winForm ...), what technologies (ADO.NET, ASMX ...) you were using, ... etc. We did this automatically and generated what I considered more harm than help (it missed things that were important and created a lot of noise.) I applied the law of relevancy and argued that we'd be better off figuring out how to leverage the user's own relevancy engine and pattern-matching ability over auto-magic guesswork. We then created a tool to help smart people, rather than a "smart" tool that gets in the way.