But don’t people only use Emacs Lisp to program Emacs?
But don’t people only use Emacs Lisp to program Emacs?
* email, slack and reddit clients * web wowser * tetris * Amazon's customer service handling https://tess.oconnor.cx/2006/03/quality-without-a-name (not sure what that says about Emacs Lisp. Or Amazon.) * PIM / OCD management software * video editing https://www.nongnu.org/gneve/ * mind control https://github.com/jonnay/mindwave-emacs * Thermonuclear word processor * vim and vi emulators, if you ever need to use it for coding * Internet of Toys https://news.ycombinator.com/item?id=29207607 * Air traffic control system https://news.ycombinator.com/item?id=33524683
While it's not clear to me how much of that backend is still in Common Lisp, when Google bought ITA Software it was primarily a Lisp shop and I understand that it largely stayed that way for at least a long time.
Yes, that's why "suspect" and "obscure": I thought that well-known system is already rewritten in Google's usual languages. Turns out it isn't publicly known (?), but plausible: Google isn't afraid to rewrite their software whenever they find legacy hard to maintain, I even heard of a project that got rewritten 5 times in ~10 years.
The ITA acquisition made them rethink that a little bit, as they noticed that all the smart people working on QPX, the kind of people Google wanted to keep more of around, really loved Lisp and had built something great in it. And then Go came on the scene and Dart and... well, so much for the four-language rule.
QPX defied bitwize's Corollary to Greenspun's Tenth Rule: "Any sufficiently complicated Common Lisp or Scheme program will eventually be rewritten in C++, Java, or Python." (These days I would add JavaScript, TypeScript, and Rust.) It was still being maintained in Lisp as of the late 2010s and probably still is today. If Google ordered a Big Rewrite of QPX in another more marketable programming language, they risk losing the good people who worked on it to begin with. So... maybe next decade.