A subjective view of Lisp

Abstract

The editor of Lisp Pointers has been asking me for a long time to write down my view of Lisp. I was even given permission to flame. This paper is the result and, naturally, is entirely my own opinion. Since the first rumors of a possible standardization of Lisp started to circulate (end of 85), many fights have taken place. They were never scientific nor technical: they were mainly commercial. Before trying to explain the state we have reached, I will try to define the ecological niche of Lisp. Lisp is old. Very old. Its remarkable conception makes it a terrific local extremum. Unlike the monsters born around the same time (Fortran, COBOL, Basic) which simply swallowed linguistic features they lacked, Lisp has evolved until it has become a sophisticated family of languages. Lisp has been the jewel--despite Joel Moses muddy analogy-of AI laboratories. Lisp has also bred a large community people for whom Lisp was the right language, enriched by a prodigious software sedimentation, which lead to exciting environments and dream hardware. But the boosting of AI which took place in the '80s is now past and the number of practising Lispers has probably gone back to what it was before. Three niches are apparent to me: hypercomplex systems, extensible systems and education.