It's clear that the software that we are using is not sufficient, and a lot of it seems to be a lack of quality control. I've been in science now for close to 10 years and the software that we use looks like it was built in 1995.
This is a serious problem with science software. The features are there (for the most part), but it seems clear that much of the software engineers are not working with the users. There is a severe lack of thought when it comes to UI and as such the learning curve is pretty steep.
With that said, my first YC application was to build software (web based) to tackle some of these issues. I'm actually looking for coders to help build this product, so if you have any interest in jumping into this field, send me an email (in profile).
I intend on applying to YC this summer and would love to work with a few programmers to give this project a real shot.