My first programming job was working in David's research group, helping transition Dasher (http://www.inference.phy.cam.ac.uk/dasher/) from a research project into one more generally useful for accessibility purposes. I still feel that that job was perhaps the most useful work I've ever done - I had the opportunity to write code that in some cases literally made it practical for people to communicate, and that (along with David's strong views on social responsibility) ended up strongly shaping my perspective on what technology's role in society should be.
I saw David only rarely after leaving the group to do my PhD, and hadn't in years before I saw him at a group reunion last month. I'm saddened that I passed on so many opportunities to learn more from him.
Regarding Dasher, David mentioned recently* that he hopes someone carries on supporting the work. It's a brilliant example of arithmetic coding in action and helps many people communicate through low bandwidth channels. Now that he's gone, I don't think any has any ownership over the project. It'll be a huge shame for it to be forgotten about. If anyone's interested in helping in someway, mention it here and I'm sure we can figure something out. Here's a quick example of it an action (https://youtu.be/ie9Se7FneXE?t=28m38s - Google Talk from 2007)
[*] David was still teaching us information theory just last month, despite going through chemo which by his admission seemed to take a lot out of him. By everyone's account he was astounding lecturer. He really cared that no one fell behind. Very sad about the news today.
I saw David only rarely after leaving the group to do my PhD, and hadn't in years before I saw him at a group reunion last month. I'm saddened that I passed on so many opportunities to learn more from him.