Sigh, Commander Taco, I'm not sure how familiar you are with basic research.
I'm not familiar with basic research in the computer sciences (although I enjoy reading the papers), but I am familiar with basic research in chemistry.
Unless you've had a sudden, brilliant insight, all research is painstakingly build on someone else's work. In fact, as a grad student you're often asked to write a review article in your area of research before you begin actual research. This gets you familiar with the lay of the land (knowledge), the direction of the field, and some of the major players in your chosen area of research. Plus, you and your research adviser might get a publication out of it.
I took a look at some of the topics mentioned on Wyvern's home page. A couple of them caught my eyes:
- Safely Composable Type-Specific Languages. Cyrus Omar, Darya Kurilova, Ligia Nistor, Benjamin Chung, Alex Potanin, and Jonathan Aldrich. Proc. European Conference on Object-Oriented Programming, 2014.
- Language-Based Architectural Control. Jonathan Aldrich, Cyrus Omar, Alex Potanin, and Du Li. In International Workshop on Aliasing, Capabilities, and Ownership (IWACO '14), 2014.
- Type-Directed, Whitespace-Delimited Parsing for Embedded DSLs. Cyrus Omar, Benjamin Chung, Darya Kurilova, Alex Potanin, and Jonathan Aldrich. In Globalization of Domain Specific Languages (GlobalDSL), 2013.
I don't know enough about the research space to know if these topics are derivative, progressive, or ground-breaking. I like the ideas that seem to be embedded in the titles, and the focus on creating safer languages / systems from an architectural point rather than brutally enforced by a compiler (see ADA).
I work in the space of HTML / CSS / Javascript / Java / SQL / Map-Reduce / Linux. It's a mess, with security, performance, reliability issues both at the boundaries and within the various containers. Something that makes life a little better in that space should be a good thing. As it is, I see front end developers not understanding back end constraints, data source developers delivering data in not-usable blobs, system admins without the necessary background to troubleshoot platform performance issues.
As a systems architect, I try to herd everything in one general direction, but to the specialists my statements can sound restrictive, demanding, and somewhat arbitrary (they're usually none of the above). If we had a common architecture / language, we might be able to converge on a better, safer application in less time, with less friction.
Don't talk to me about UML. I like it, but I don't know many who do or use it much past the design phase (if that).
In short, it looks interesting and ambitious. I'll be interested in seeing how far this gets.
--- end Sunday night rant