criteria

Spend only a little time each time you work on it.

Decouple the effort from any other existing commitments and projects.

Seek potential for useful value in about three years. The project works when having one of something has little value, but having 1000 or 5000 of them has plenty of value.  Realize that it might take more time than you plan to make it happen.  

Note things that are happening across a border or around the world, not just here.

Identify sources of leverage, where a complex current process can be simplified.

Consider that tools tend to attract tool-wielding people, and problems that require tools to attack.

note: edited 10/27/10 to make it make a little more sense.

Comments

[Edward Vielmetti]

Curiously, this is a type of project design which is difficult to deliberately start. You just kind of have to notice after a while that you have collected 100 or 200 of something, and that you're barely started. Not every effort is addressable in this fashion.

[Edward Vielmetti]

Also curiously, this is a project design that needs to deal both with short bursts of intense activity and perhaps long times where there is nothing going on at all, as other plans that are working in parallel intrude.

[Edward Vielmetti]

Notice a conflict between tool building and doing. Are you building a tool for laying out pages in a book, or are you writing a book? Are you building a collection of tools, or are you using those tools?

Perhaps more importantly, does the value accrue to you or to the system whose tools you are using? That "free hosting service" has a business model of its own.

[Edward Vielmetti]

The reality check is that "2500 of them" is 7 years of producing something daily. That's a lot of days.