I'm subscribed to a lot of RSS feeds. One of them is from Somasegar's weblog. He's head of Microsoft's Developer Division. His recent article got my attention. He writes

"As a development organization, it is important to make sure that a portion of the team’s bandwidth is allocated for incubation or “long-lead” innovation which we expect to take to market in a subsequent product cycle.  This means that a business has to plan resource allocation in the following 3 buckets

· Making current customers successful (servicing)

· Working on the next version of the product

· Incubation"

He's absolutely right. If you always work on maintenance and the upcoming version only and don't allow your development team time for learning and trying new stuff, "nurturing incubation" as Somasegar calls it, you won't get real innovations, your staff's knowledge will get stale and motivation will suffer. It's not enough hoping that everyone learns new technologies or improves their know-how during spare time, just because it's in their own interest. 

"Resourcing incubation has to be a part of planning – otherwise the natural tendency is to suck up all the resources for the critical activity of the day."

And it takes people with the right mindset and capabilities to profit from such incubation time. Somebody who does programming just as a 9 to 5 job won't know what to do with incubation time.