Ok the other hand, a good lisp implementation of neural nets would offer a specific set of macros dedicated to matrix operations. This set of macros is, in the

Ok the other hand, a good lisp implementation of neural nets would offer a specific set of macros dedicated to matrix operations. This set of macros is, in the lisp point of view, a domain specific language. So, you still have 2 languages.