The latest Sane book is uploaded, a work in progress.
The major development is the start of parsing. The internal
data structures are the target and they have already been
mostly developed as the system is being written "middle-out".
That is, the intermediate representation was the first thing
The Axiom parser, developed by William Burge, has a lot of
permutations of syntax. The new parser will likely restrict the
syntax. In addition, the new parser will have to handle logic
forms which have yet to be chosen. There are several logic
languages with various tradeoffs. Some of the possible
choices imply a lot of hidden machinery, such as the ability
to make arguments optional. That implies a prolog-like
resolution algorithm which adds a lot of additional overhead.
Since we're striving for simple and maintainable code as
well as clear end-user language features it is not clear that
this overhead is beneficial.