Editing Chaitin's construction (section)

== PROBLEM with the approach below ==

The approach below fails to take into account that a halting probability is only (provably) random if the programming language used is additively optimal. That is, it must provide access to raw binary data without overhead. Thus, a binary code for Combinatory Logic or Lambda Calculus does not suffice. It's also necessary to apply the encoded term on the binary data that follows the encoded term, as is done for example in Section 4.1 of [https://tromp.github.io/cl/LC.pdf Tromp's paper].