This just reminds me of how many times I've been told things like "you can't use that java code that is open source and working and does exactly what we want - we're a python shop!" and other such nonsenses.
I'm always amazed at how people have strong feelings for the tools/languages, rather than what they do and how they can help you. Even in this example, the Google person didn't even ask what they wanted to use Lisp for.
People I guess just really like rewriting software in different languages?
Ugh, you haven't experience pain until you've worked on a project which has a bunch of different languages because Johnny wanted to use Scala, Freddy loved PERL and Sally was learning Elixir on the side. They all left and the new team has to figure out how all these stupid languages even build.
I didn't say "let everyone go off and learn a different language and develop new code in it," obviously that is ridiculous, and would get out of control as you said.
What I did say was there's a working piece of code out there that someone else has already spent time getting to work, but we can't use it because of the language it's written in. For example, I've had problems getting people to use Jenkins because it's written in java. Not that they were trying to write plugins or do anything internal to Jenkins, they just didn't like it because the language it was written in.
As an engineering manager, you could support every team or even every individual using whatever they like - but you then have other tradeoffs to make.
It's now not possible to run security tooling over everything or use a common artifact repository or follow common coding guidelines or linter settings and you'll get islands of language culture and endless internal reinvention and rewriting.
For me it makes lots of sense, it is not "nonsense".
Java is a gigantic dependency. Any project that depends on it adds all this dependency with it. In many cases this dependency could be bad or toxic for the project.
It depends of what you do for a living. Personally I will never let anyone include java as a dependency in my company's projects, because for what we do, java is slow and linked to Oracle.
Lisp is a different story. You can create c ,Python, Lua,c++ or whatever code from Lisp easily.
You will see it defended a lot on Hacker News, due to the disproportionate level of PL enthusiasts, and the Paul Graham thing. But yes, you're right that in reality, lisps have incredibly limited popularity. (You've met one more full-time Lisper than me and I've been in the industry for 10 years).
In my current and in my last 2 jobs I've been using a Lisp (Clojure). Since learning Clojure and functional programming in general, I made a conscious choice to seek it out for the last two positions (a small startup and a huge government project) and in this current one (university research) I'm starting two projects from scratch and get to decide myself which language to use, so of course I'm gonna choose the one I enjoy the most and will be more productive in. It's hard to go back to imperative/oop.
What for? I mean, I have relative freedom at work and could probably use Lisp on my own tools for my own job, but I’ve never really considered it (I use Haskell and Python instead).
I'm always amazed at how people have strong feelings for the tools/languages, rather than what they do and how they can help you. Even in this example, the Google person didn't even ask what they wanted to use Lisp for.
People I guess just really like rewriting software in different languages?