Re:The 8000th Busy Beaver number eludes ZF set theory (2016)
by lmm on Friday May 04, 2018 @05:55AM
(#16993360)
Sounds pretty straightforward. This is a normal way of working in functional languages: you write your business logic in a high-level representation (aka an "embedded DSL") and a succession of interpreters that interpret slightly higher-level representations into slightly lower-level representations, until you get down to one that you can implement directly. It's a nice structure for testing because you can test each interpreter separately on each of its instructions and then the composition behaviour is guaranteed.
Re:The 8000th Busy Beaver number eludes ZF set theory (2016)
by xyzzyz on Thursday May 03, 2018 @04:39PM
(#16989723)
Sounds like an interesting assignment at a good undergrad compilers course.
He then managed to test it on a simple statement about square numbers.
Sounds unimaginably complex.
How many people could help you write unit tests for such a thing?