Because as 'klibertp wrote, they're two vastly different languages, with different design decisions.
The whole thread is about primarily Common Lisp; people who enjoy the design choices made by the CL standardization committee and subsequent evolution of the language's ecosystem would most likely want to know that there exists a proper CL implementation for JVM.
but the person i replied to replied to a commenter lamenting the fact that there was nothing but java and c style development jobs instead of lisp inspired jobs. they weren't lamenting the lack of jobs specific to just common lisp. and to me, it doesn't make sense to "sneak in" an esoteric implementation of lisp on the jvm when clojure already exists. what makes people think that would be easier to sell at a place using java heavily?
I wouldn't call ABCL esoteric; it's as rightful JVM citizen as Clojure or Scala. That said, I agree it's much, much less popular in the JVM world, so it could be a harder sell.
of course they're different. i didn't say or imply otherwise. the original parent mentioned both racket and common lisp, again different languages, so it was clear they weren't looking just for common lisp and were more generally wanting for some lisp or scheme type of job. and i was asking a legitimate and honest question, as i see ABCL to be a much more esoteric and hard sell than clojure to a place heavy on java development.
You asked, "why would one use [ABCL]?" What I implied in my response (sorry for not stating it directly) was that one may prefer the particular feature set of CL over Clojure's.
As for being a hard sell - my personal experience is that if you're in a position that allows you to choose your implementation language, in most cases you can choose Brainfuck and nobody will care. As an example, I introduced F#, Erlang, Elixir and a bit of Prolog to a Python shop that way: all that mattered in these cases was whether the code worked, and it did, so no one complained. The Erlang microservice actually outlived the project it was written for by a year and a half.
On the other hand (still, in my experience) if you're not in a position where you can choose your language you're basically screwed, and should stop trying to convince your team to switch to something else. You have almost no chance of succeeding and a high chance of disrupting teamwork by pushing your opinions on others.
So I was thinking about the former situation, where you have some degree of freedom of choice and can decide the language on technical merits alone.
The whole thread is about primarily Common Lisp; people who enjoy the design choices made by the CL standardization committee and subsequent evolution of the language's ecosystem would most likely want to know that there exists a proper CL implementation for JVM.