I wrote my master's thesis in overleaf about a year ago, because my laptop was starting to break down and I needed a solution that allowed me to easily switch between university computers without having to constantly set-up my workspace on each individual PC. That was before the rich content editor was implemented, but I loved the experience nonetheless. The whole thesis was, next to my research, an exercise in cloud-based software. VM's on Cloud9, Code repo on github and my text editor on overleaf. Thanks a lot for making it available to a poor student like me!

Thanks, that's great to hear! :) I'm impressed you used your thesis as a testbed for cloud-based software, hope it didn't cause any/too much additional stress! If you generated any datasets during your work, you may also want to check out figshare[1] as a handy data repository.

[1] http://figshare.com/

(Note - we work in the same office as figshare, so this is probably a biased recommendation!)

Thanks for the link! My thesis was basically a data analysis project, improving on work my supervisor did prior to me starting my thesis. Figshare is good to keep in mind, but I don't think I'd have had much use for it, as the whole point of my thesis was to get certain figures within a reasonable time.

I actually had quite a lot of fun figuring out how to get everything working in the cloud. Set up my home desktop as my personal computer "cluster" (i5 processor, so a cluster of 4 cores!), I find it hard to work in an environment that I tend to do other stuff besides work in. Derived a small representative dataset from the massive one I had to quickly iterate over my analysis scripts in Cloud9 (while having github take care of my revision history). Then ssh'ing into my desktop PC whenever I needed to run the scripts on the full data set.

I considered adding a section on open science and cloud-based software to my thesis, but in the end decided not to due to a lack of time. Open science is based on open source principles, where everything that can be shared should be shared. The Cloud VM angle is quite interesting in this sense, as being able to share the machines that software for a research project was ran on is often quite difficult (or nigh impossible), and a cloud-based VM seems to bridge this gap somewhat. In theory, anyone could log-in to the VM and play around in the workspace I set up for my thesis project, not that many would want to I suppose.