A state of the art compiler [1] in quantum computing has not been rewritten in its 8 years of existence. This is accompanied by a quantum computer simulator that runs on the world's biggest supercomputers. [2] Multiple quantum computing companies use Common Lisp, and their software was written in a fraction of the time and resources taken for competing companies to do the same in either C++ or Python.
Things like Coalton [3] really make Lisp stand out. What other language is able to add Haskell-quality types to their language?
Lisp is not popular, so it doesn't have as much labor going into it, but for the work that does, remarkable results are achieved. Lisp does have a lot of the dubiously claimed power, but that power still requires reasonable engineering work to achieve—work that a lot of hobbyists are unwilling to do in their free 4 hours on the weekend.
Lisp software like QUILC (linked) and Python software like Qiskit.
They have the role of taking general quantum programs (usually called "quantum circuits"), and translating+optimizing them for a particular quantum computer architecture (i.e., one that might only have a small number of supported quantum operations).
It's analogous to a language like C having a compiler (written in Lisp or Python) for ARM and x86.
I hesitate to give quantitative estimates without citations, but to throw a bone, something like: <2 years for Lisp by a small team of 4 relative newcomers. >4 years for Python by a larger team of >10 expert people. The software still does not have feature or performance parity.
"Until now"? Do you really think that this is the first time a Lisp has been used to write an application? Is that because this is the first time you're hearing about it?
You'll notice that most of these aren't user-facing applications or they're extremely specific to a particular industry. Common Lisp itself was standardized largely due to the demand of the US Department of Defense so they would have a single, combined dialect of Lisp for DoD/DARPA projects. That's the sort of language this has always been (entirely out-of-sight, and also for most of its life it was prohibitively expensive). It's similar to Ada, in that sense.
Python obviously does run on those same computers, yes, but most Common Lisp implementations have the benefit of compiling to machine code, while Python is interpreted.
Jeez lispers have the most enormous chip on their shoulders. You really don't see the irony in a comment that points to a compiler for a lang for a quantum computer as proof lisp is used to build important things?
> US Department of Defense so they would have a single, combined dialect of Lisp for DoD/DARPA projects.
I've worked on DoD projects (through NVESD) and this is like almost bald-faced lie. While I can believe DARPA has/had some connection to lisp (because of early expert systems) it's nearly impossible to believe the actual military has any lisp code anywhere. To with Ctrl+f here
> Today, Ada is the most commonly used language for mission-critical defense software, which includes weapon systems and performance-critical command, control, communications, and intelligence (C3I) systems. DOD's inventory contains nearly 50 million lines of Ada code in these applications
Not sure if it's "impossible to believe," but I suppose everyone else is lying too.
I searched through dtic.mil w/ Google and there's a great amount of reports that detail DoD-funded research and projects that involved Common Lisp. This one is the most unambiguous:
https://apps.dtic.mil/sti/tr/pdf/ADA303486.pdf
"The Composite Warfare Model (CWM), a Navy simulation used successfully by the Space and Naval Warfare Systems Command (SPAWAR) and others, uses a fairly radical decomposition in which most attributes and methods are in their own class. Objects may have hundreds of super- classes. Written in Common Lisp, CWM is "linguistically advantaged" for this approach; ..."
An amount of those could've never been used (some are just proposals, and I'm not reading more or searching to see if they went anywhere) but some of them certainly were. It's mostly simulation and planning software, perhaps not mission-critical (I'd think garbage collection would be an obstacle there).
> While I can believe DARPA has/had some connection to lisp (because of early expert systems) it's nearly impossible to believe the actual military has any lisp code anywhere.
It may be hard to believe, but just 7 or 8 years ago while working for DOD my colleagues the next aisle over were maintaining Lisp systems (the whole org was focused on maintaining systems, so I can offer little insight into new work, but probably not much in Lisp). So perhaps it's impossible to believe, but it's true.
Many of my colleagues there (who had been DOD employees for decades) got their start in the "AI Lab" there doing Lisp programming in the 80s.
Things like Coalton [3] really make Lisp stand out. What other language is able to add Haskell-quality types to their language?
Lisp is not popular, so it doesn't have as much labor going into it, but for the work that does, remarkable results are achieved. Lisp does have a lot of the dubiously claimed power, but that power still requires reasonable engineering work to achieve—work that a lot of hobbyists are unwilling to do in their free 4 hours on the weekend.
[1] https://github.com/quil-lang/quilc
[2] https://github.com/quil-lang/qvm
[3] https://github.com/coalton-lang/coalton