I remember reading the Symbolics manuals from bitsavers and how I have been impressed by clarity and unity of the language, and how they incorporated CL into Zetalisp just by moving their own stuff into packages.
At least in the documentation everything looks wonderful. Compared to Symbolics CL looks like a mess at least from reading the books.
The major goal of Common Lisp was commonality, and that means incorporating forms from various dialects into one and to maintain compatibility, like you said. That goal has been accomplished.
I think that carefully designed dialects with emphasis on right principles and attention to details leading by people like David Moon would be aesthetically better than a good kitchen sink. It is no coincidence that Common Lisp took most of stuff from Zetalisp, according to CLtL2.
I think it is a general heuristic that a small group of disciplined devoties would produce better artifact than an vast assembly of... general public or passionate and productive but ignorant individual. Think of Python3 vs Ruby, Go vs early C++ or Scheme vs CL. Attention to details and perfectionism works - look at Haskell syntax and prelude. (All this just to illustrate validity of my heuristic).
> The major goal of Common Lisp was commonality, and that means incorporating forms from various dialects into one and to maintain compatibility, like you said. That goal has been accomplished.
That's misleading. The 'various dialects' were mostly three very similar successor implementations of Maclisp.
Common Lisp was designed as a single successor language to Maclisp. There were successors of Maclisp for different computers under development: NIL (on a super computer and the VAX), Spice Lisp (for a new breed of workstations), Lisp Machine Lisp (developed for MIT's Lisp Machine systems), ... We are talking about new implementations of basically very similar languages, with Lisp Machine Lisp as the major influence.
DARPA wanted that these and future successor implementations stayed compatible. Not every Lisp application for the military should come with its own incompatible implementation of the language. NIL then looked a lot like CL. Spice Lisp evolved into CMUCL. Lisp Machines then got a CL implementation integrated with Lisp Machine Lisp.
From CLTL:
> Common Lisp originated in an attempt to focus the work of several implementation groups, each of which was constructing successor implementations of MacLisp for different computers.
Common Lisp was not designed to be compatible with other Lisp dialects like Interlisp, Scheme, Standard Lisp, ...
At least in the documentation everything looks wonderful. Compared to Symbolics CL looks like a mess at least from reading the books.