Garbage Collection Based on Linear Type System
Thank you for your thoughts on the GC paper. I haven’t seen any
follow-ups, and there were quite a few serious todo’s in the ‘further
work’ section of the paper, so it seems like perhaps it didn’t have
enough vim and vigor in it. Still, I’m happy there are people trying
out crazy ideas!
Makes sense? I am not sure. It is an interesting idea largely based
Andrew Tolmach’s work on tag-free GC. If you want me to bet, my bet
is that it is unlikely to work. There are many reasons. The need for type
information at run-time may be too much of a burden. How fast can such
type information be inferred at run-time? Also, issues like handling data
generated by foreign code are not yet touched.On Wednesday, October 21, 2015 at 8:52:21 PM UTC-4, Raoul Duke wrote: