As a cheap experiment with potential, i think this idea is a good one. In the longer run, I think we should consider some broader issues:

1. Maybe np, the broader nonprofit sector, would be worth considering, rather than couching things in nptech. imo, there's already a pronounced tech bias in delicious.
2. limitations of delicious as a tool. What are these and how will these affect the success of this project? multi word taxonomy entries aren't possible in a consistent manner, the 'earlier/later' dichotomy in delicious may not be an appropriate filter on quality and relevance. what's the potential for gaming the system by a small number of people with uncommon interests?
3. audience. I think we should be clear about the intended audience. IMO, i think that the nptech taxonomy is intended primarily for use by people who work in this field, and secondly by np staff. Maybe this isn't right.
4. goals. a few goal suggestions:

• to have a well-organized, self-maintaining bookmark library that we can share.
• to discern a taxonomy by observing how nptech's classify resources
• to get a sense of the quantity & diversity of web resources within this nptech field.