4 ms·
One thing I think will eventually come out of linear logic is that in classical linear logic there's no intrinsic division between inputs in outputs. In LL it's
by danharaj 8y ago
One thing I think will eventually come out of linear logic is that in classical linear logic there's no intrinsic division between inputs in outputs. In LL it's the case that implication, which corresponds to functions in programming, is decomposed into a linear implication, which is a pure application, and an exponential which handles copying. But even further than that, in classic linear logic there is a symmetric version of implication called par. A -o B is the same as Not A `par` B.
It's this symmetry that intrigues me, because it points to a way of programming beyond functional programming that still maintains its fundamental good properties.
- YorkshireSeason 8y agoI recommend looking at fusion calculi, which are based on the idea of input/output being symmetric. They were invented independently by B. Victor together with J. Parrow, and at the same time by Y. Fu. See [1] for some links. NB this is an old page, and I don't think anyone currently works on this. When it was first released, Microsoft's BizTalk was based on a generalisation of Fusions. (L. Wischik, the author of [1] what on the BizTalk team.) [1] http://www.wischik.com/lu/research/fusions.html http://www.wischik.com/lu/research/fusions.html