Most of the problems with medical software is the doctors and medical administrators themselves. They get so used to be worshiped as Gods That Save Lives, they assume that they Know Everything About Everything. I've seen more software and business talent quit the medical software field than any other....
I quit the medical field because it's incredibly hard work and very difficult to succeed because of the additional hurdles (regulation, entrenched interests) unless you are a large corporation. And large corporations suck.
My friend tells the story of being in one meeting with the various doctors, and one of the elder ones made a snide comment about another doctor 'not being here long enough to understand how things are done in this place' - to which the reply was 'I've been here for 17 years'.
Myself personally, I worked as a neurology tech for four years, and the doctors I worked with were a broad mix. Some of them were the sweetest people you could hope to meet. Others were such money-grubbers they directly said to the pharmaceutical reps "don't give us 'gifts', just give us the money directly". In that broad mix were doctors so awesome that they write their own mouse drivers, to doctors that behaved just as xradionut mentioned.
Doctors are humans too, there are a lot of them, and they run the same kinds of personalities as other humans.
My doctor had to sell her private practice and join a medical business group, just to keep up with the new insurance/government regulations. Healthcare in the US is insane...
Seems like the perfect situation for making a software workflow and simply not permitting information to be input into Field X. Etc. Anyway, in this case, I think the software could be improved (modulo my slow learning).
Now, in general, I think that software can help us make the bureaucracy / regulation more tolerable by helping to convert medical intent into both medical action and medicolegal documentation.
Anyway, your point is very well taken.
Are you sure who was who? :)
I would bet that there is a gigantic initial market for the cure for cancer.
The objective requirement: in the success-case the project must have a large impact on the community (in Hamming's case, physicists; for strtups, an initial market / society at large)
The subjective requirement: the researcher must have "an attack" [0], i.e. a credible approach to a solution that others probably have not tried yet.
For the vast majority of SF Bay Area founders of tiny software companies, they have no "attack" on cancer.
This isn't exactly true but the truth that it hints at is largely about mass standardisation, accountability and global purchasing rules for large organisations.
The same applies at telcos, banks etc. You can turn an oil tanker around quickly.
Meanwhile, your doctor in the hospital is stuck using some 1980's software, which drains the very life-force out of him by the amount of pain it inflicts with every use, but you'll never need a doctor anyway, not with this healthy-eating/exercising app, plus it's too hard to work on that with the regulations and all, and the carpeting in your apartment is moldy.
[0] https://news.ycombinator.com/item?id=6106528
[1] http://www.youtube.com/watch?v=guf0kNV7D1w
ps. On an honest note, I've been working both on apps of the "useless" type, and spend most of my time on working on "useful" medical apps, and fuck if it's not easy to see why no one thinks these areas "sexy", so bogged down in regulation and old entrenched players. But I still think, at least in the case of these super incubators, who have the connections, who have the access to the financing, the focus should be on something more than what it is right now.