The problem is people working in natural language processing (NLP)
consider operator parsing a rather trivial problem. Since many NLP
problems are solved with a lexicon, which is very close to an operator table.
When I mention Alain Colmerauer I was not talking about particular syntax
of a Prolog system. But about the parsing mechanism of definite clause
grammar (DCG), because the OP asked which parsing algorithms to use.
Amazingly DCG was already used by Alain Colmerauer in reference I gave.
When translated to SWI-Prolog noadays, his examples would still work, one
needs to find a way to map “suprimes” and “retarder” though. And when I look
at the date of the report it is from 1972. While Pratt promoted BNF in 1973
comming from a CGOL project. The difference between the Alain Colmerauer
report, and the Pratt report, the later highlights more parsing artificially created
languages, such as programming languages, and their operator problematic.
While the former demonstrated more a natural language processing
use case, like a query answering system, in the various traditions of
