I haven't written much about my progress on Higgs recently. This is because I've been busy writing and submitting conference papers. My ECOOP paper about basic block versioning has been accepted, and I just submitted a new paper about typed object shapes to DLS 2015. In July, I'll be presenting my research at ECOOP in Prague, and then I'll begin work on a third paper about interprocedural basic block versioning, which will complete my PhD work.
My thesis advisor has suggested that I construct my thesis as a series of conference papers. I already have two papers published about my work with JavaScript. The paper I just submitted and the next one will also be included, for a total of four papers. If everything goes well, I should be submitting my thesis in December 2015 and defending it in February 2016. After nearly 6 years in the PhD program, I can finally see the light at the end of the tunnel!
The unfortunate thing, for me, is that coding and working on Higgs is the part of my research that I enjoy the most (what can I say, I like building systems), and I haven't had time to do much coding recently. Fortunately, there are a few more improvements to Higgs that I'll implement for my next paper. I'll be improving the calling convention, which should increase the performance of function calls quite a bit. I also have plans to improve the performance of global variable accesses even further.