Smarter Messaging [Mike Jones, UK]

I can’t remember where I heard or read the following thought (roughly speaking):

You can’t create tools that don’t reward the user if they commit anything less than 100% because they won’t get “popular”

The example they gave was of a calendar application which if you don’t put all your appointments in it you can not “trust” it. I’ve been trying make sure I apply this to system design I’m thinking about which has multiple points of data entry and where there are lots of ties and links between those individual silos. If the above statement is true then the silos have to be individually valueable and the value that is gleaned by linking the them together must not be a deciding factor in implementing it.

However, if I can make it so the “extra mile” required to do the inferred linking is so easy and valueable to the consumer of that data (and not just their boss or co-worker) then all parties will win. An example of this would be tagging on del.icio.us, the social bookmarking site. On the site the more keywords (tags) you associate with a bookmark you add there, the more likely you are to find it when you go looking for it. The flip-side is if I tag something “IM” and someone else “InstantMessaging” then the same thing can be found via many vectors. Of course del.icio.us takes it to another level by analysing its data and finding closely related “stuff”, which I have to think about for this system.