Ask HN: Who is using Common Lisp?
What/why?
56 comments
At the University of Washington Medical Center (UWMC) we
are using a DICOM server written in Common Lisp [1]. DICOM is a network protocol used for medical images and other medical data. This
was originally developed as part of the Prism radiation therapy treatment planning system [2][3], which was completely written in Common Lisp, and was used as the sole
planning system at UWMC for many years.
1. https://dl.acm.org/citation.cfm?id=963921
2. https://www.ischool.utexas.edu/~ml/papers/lisp40yrs.pdf
3. https://www.redjournal.org/article/S0360-3016(96)00322-7/abs...
1. https://dl.acm.org/citation.cfm?id=963921
2. https://www.ischool.utexas.edu/~ml/papers/lisp40yrs.pdf
3. https://www.redjournal.org/article/S0360-3016(96)00322-7/abs...
Rigetti Computing [1] uses it for a variety of their projects, including their high-performance quantum computer simulator called the QVM [2], and their quantum compiler.
Some of their use of Lisp has been presented in more detail at the European Lisp Symposium. You can see it in the Cutting Edge section of the proceedings [3].
[1] https://www.rigetti.com
[2] https://youtu.be/f9vRcSAneiw
[3] https://www.european-lisp-symposium.org/static/proceedings/2...
Some of their use of Lisp has been presented in more detail at the European Lisp Symposium. You can see it in the Cutting Edge section of the proceedings [3].
[1] https://www.rigetti.com
[2] https://youtu.be/f9vRcSAneiw
[3] https://www.european-lisp-symposium.org/static/proceedings/2...
Then you should mention Grammarly, with about 10 million users per day. Orbitz has similar numbers.
A month ago I built an entire working MVP for a cloud-based biomedical knowledge management, discussion, and decision-making system, entirely in CL (ccl), using raw HTML, down to the metal, in a week. And we went from MVP to mid-VP (near beta) in another three days, and were running live user experiments at the end of that week. ~2000 lines of CL. One.5 weeks' work. MVP NLP and ontology management, discussion management, role management.
Yes, I'm for hire! :-)
ps. That was a month ago. What I've done in the meantime, aside from wind surf, is build the AI underneath that -- another 2000 lines of working CL. Lisp Rules!
Yes, I'm for hire! :-)
ps. That was a month ago. What I've done in the meantime, aside from wind surf, is build the AI underneath that -- another 2000 lines of working CL. Lisp Rules!
How many years of experience do you have? And is there anywhere I can checkout the product or learn more about it?
Sounds interesting.
Sounds interesting.
What is the best way to contact you?
Post a gig (free!) on SFCL...uh, that's SF CraigsList (not SF Common Lisp :-) If it looks interesting, I'll answer it. If I don't, you're likely to get other reasonable folks answering it. (Only reasonable folks answer Lisp job ads, because only reasonable folks know Lisp; It's not like Python where kids who have one high school python course are applying for jobs!)
Haha...a friend of mine, reading the above over my shoulder, reminded me that there are, like, two complete a-holes over on comp.lang.lisp who long ago drove everyone else away. (Usenet is pretty much dead anyway, but these folks hastened the early demise of c.l.l as a useful forum, even before the rest of usenet folded to Stack Overflow.)
ta46854884(1)
SISCOG [1] uses on all their products. Actually they use Allegro CL from Franz. [2]
The main reason (for still using it) being that it all started out as an end of college project, by the two CEOs, around Artificial Intelligence proof-of-concept-general-tools (so to say) until the big breakthrough for danish railway company, NS, which was impressed with their fleet scheduling project. This adventure began in the late 80's and they were using Lisp Machines!
I suppose the reason for still using Lisp to be a mix of i. there is no technological necessity to change; ii. the codebase of existing products is too big to translate; iii. CTOs don't have enough knowledge on other technologies to evaluate the feasibility of a change; iv. many of the developers actually love Lisp :-).
Most of the support technology is also done in-house and they don't tend to use many external libraries.
1. http://siscog.eu/
2. https://franz.com/products/allegro-common-lisp/
The main reason (for still using it) being that it all started out as an end of college project, by the two CEOs, around Artificial Intelligence proof-of-concept-general-tools (so to say) until the big breakthrough for danish railway company, NS, which was impressed with their fleet scheduling project. This adventure began in the late 80's and they were using Lisp Machines!
I suppose the reason for still using Lisp to be a mix of i. there is no technological necessity to change; ii. the codebase of existing products is too big to translate; iii. CTOs don't have enough knowledge on other technologies to evaluate the feasibility of a change; iv. many of the developers actually love Lisp :-).
Most of the support technology is also done in-house and they don't tend to use many external libraries.
1. http://siscog.eu/
2. https://franz.com/products/allegro-common-lisp/
Planning and Scheduling is a domain where Lisp has historically seen a lot of applications. For example the planner for the use of the Hubble Space Telescope (and many other telescopes in space and on earth) was originally developed on TI Lisp Machines and later ported to Common Lisp on other platforms. Generally there were a bunch of Lisp-based planners for space operations (like payload planning for Space Shuttles or planner for mars operations). There were a bunch of Lisp-based planner application in military - with the most famous being the DART system - which was used for planning the logistics for the first gulf war and saved more money that the US had ever invested in AI research up to that point in the past 30 years. With field-deployed Common Lisp applications in the middle-east, talking to remote servers. https://en.wikipedia.org/wiki/Dynamic_Analysis_and_Replannin...
Indeed. As for space exploration goes I don't remember if it was The Martian or Interstellar which mentions Nasa's Jet Propulsion Lab (JPL). And from there I found this nice article about Lisping at JPL, by Ron Garret [1]. I think it's an enjoyable reading.
1. http://flownet.com/gat/jpl-lisp.html
1. http://flownet.com/gat/jpl-lisp.html
Wow!!! I had no idea. Thanks for the link. :-)
I want to echo Grue3[1] that the tooling around common lisp is amazing; some tools I've used are better at one level of abstraction, but none are as good all up and down the levels of abstraction. Very few languages (smalltalk comes to mind) have the culture of caring about tools.
A lot of other features that are great about lisp have become mainstream in the 15 years since I learned lisp.
Or to put it another way, python feels like a slow lisp implementation with a friendlier syntax[2], a large standard library, and poor metaprogramming support[3].
1: https://news.ycombinator.com/item?id=17617889
2: Some lispers will be annoyed by me calling python's syntax "friendlier" than lisps, but it seems to me to be true, inasmuch as this can be objective.
3: One major reason for lisp's syntax has persisted is how easy it makes metaprogramming. Once you've learned the basics of lisp, I can explain to you how Common Lisp's macro system works in under 5 minutes. The ratio of power/simplicity of it is really astounding.
A lot of other features that are great about lisp have become mainstream in the 15 years since I learned lisp.
Or to put it another way, python feels like a slow lisp implementation with a friendlier syntax[2], a large standard library, and poor metaprogramming support[3].
1: https://news.ycombinator.com/item?id=17617889
2: Some lispers will be annoyed by me calling python's syntax "friendlier" than lisps, but it seems to me to be true, inasmuch as this can be objective.
3: One major reason for lisp's syntax has persisted is how easy it makes metaprogramming. Once you've learned the basics of lisp, I can explain to you how Common Lisp's macro system works in under 5 minutes. The ratio of power/simplicity of it is really astounding.
python friendlier? just kinda different, i'd say, like algol-ier
I'm planning to write a game for Ludum Dare 42 using SBCL and clopengl. I wrote a simple Asteroids game way back with SDL (https://github.com/andyhd/asteroids), but I haven't had much opportunity to use CL since then.
I find Lisp is the most enjoyable language to write in, because it has such minimal syntax, but is incredibly flexible.
I'm using Common Lisp for all my personal projects (https://ichi.moe is the biggest one). I find it the easiest language to develop in, because of the tooling (SLIME), language features (CLOS, macros), and it fits my mental model the best. Granted I've been using it for a very long time, so maybe I'm just more familiar with it.
I have no real opinion on the tech stack (beyond being glad that it works for you), but this is very cool, and is something I've looked for in the past, unsuccessfully.
For something a bit more mature, you might also like https://jisho.org , which uses the venerable MeCab [1] for morphological analysis, instead of the apparently homegrown system of ichi.moe, and also incorporates a lot of other resources.
[1] http://taku910.github.io/mecab/ , also available as a Debian package
[1] http://taku910.github.io/mecab/ , also available as a Debian package
Except Jisho obviously produces worse results on pretty much any non-trivial sentence. Sure, I wrote my own segmenter because everyone else's wasn't good enough, that's a feature not a downside.
Can you give an example?
Just a few deficiencies that I found:
1. Jisho can't really do conjugations. It only says "X could possibly be an inflection of Y with these forms" based on the ending of the verb. For example 邪推してしまう becomes
邪推してしまう could be an inflection of 邪推する, with these forms: Te-form. It is a connective form of the verb, usually called the gerund.
As you can see it only saw "邪推して" and ignored all the rest. ichi.moe always fully parses each conjugation.
2. Jisho can't handle when words are spelled in kana, unless the word is specifically marked as "usually kana". For example when parsing a sentence "ヤバそうなはなしを聞いた", the word "ヤバそう" is not recognized at all, and "なはなし" is parsed as "なは"+"なし". ichi.moe can often handle such spellings.
3. It splits expressions like "いざとなったら" in "いざとなったら勇気が無くなった" (despite the word obviously being present in their dictionary) and the meaning of the expression can be hard to guess from the components.
Considering their segmenter apparently last updated in 2013 there's little hope this will ever improve.
1. Jisho can't really do conjugations. It only says "X could possibly be an inflection of Y with these forms" based on the ending of the verb. For example 邪推してしまう becomes
邪推してしまう could be an inflection of 邪推する, with these forms: Te-form. It is a connective form of the verb, usually called the gerund.
As you can see it only saw "邪推して" and ignored all the rest. ichi.moe always fully parses each conjugation.
2. Jisho can't handle when words are spelled in kana, unless the word is specifically marked as "usually kana". For example when parsing a sentence "ヤバそうなはなしを聞いた", the word "ヤバそう" is not recognized at all, and "なはなし" is parsed as "なは"+"なし". ichi.moe can often handle such spellings.
3. It splits expressions like "いざとなったら" in "いざとなったら勇気が無くなった" (despite the word obviously being present in their dictionary) and the meaning of the expression can be hard to guess from the components.
Considering their segmenter apparently last updated in 2013 there's little hope this will ever improve.
When I do echo '邪推してしまう' | mecab I get
The other problems do seem to be related to the segmentation approach MeCab takes. I've usually found it to be satisfactory, but maybe I just haven't been looking hard enough for errors. I might try and see how well your Ichiran works as a replacement. Thanks!
邪推 名詞,サ変名詞,*,*,邪推,じゃすい,代表表記:邪推/じゃすい カテゴリ:抽象物
して 動詞,*,サ変動詞,タ系連用テ形,する,して,代表表記:する/する 付属動詞候補（基本） 自他動詞:自:成る/なる
しまう 接尾辞,動詞性接尾辞,子音動詞ワ行,基本形,しまう,しまう,連語
EOS
The other problems do seem to be related to the segmentation approach MeCab takes. I've usually found it to be satisfactory, but maybe I just haven't been looking hard enough for errors. I might try and see how well your Ichiran works as a replacement. Thanks!
Which implementation do you use?
SBCL aka the best one. I develop on both Windows and Linux machines, depending on circumstances and it works pretty well on both.
Better than Allegro or LispWorks?
For me the biggest pain is that there are plenty of libraries that simply don't "quickload" in LispWorks. I've never had that issue with SBCL, which about 70% of the community seems to use, followed by CCL.
No surprise considering how gimped the personal edition is (which is even one major version behind). I wish they had a sort of revenue restricted Community Edition just like embarcadero is doing now. I'm old and I don't mind paying for tooling, but how many are dropping $750 on the Hobbyist edition?
No surprise considering how gimped the personal edition is (which is even one major version behind). I wish they had a sort of revenue restricted Community Edition just like embarcadero is doing now. I'm old and I don't mind paying for tooling, but how many are dropping $750 on the Hobbyist edition?
I see, yes you are right about the price.
I tend to call the attention for these Lisps, just because it seems a large amount of developers aren't aware what they are actually loosing in productivity by sticking with their free (gratis) tools.
Embarcadero is another good example, given what Delphi and C++ Builder were already capable of. Although here they have themselves to blame how Borland lost their hobby market ethos.
I tend to call the attention for these Lisps, just because it seems a large amount of developers aren't aware what they are actually loosing in productivity by sticking with their free (gratis) tools.
Embarcadero is another good example, given what Delphi and C++ Builder were already capable of. Although here they have themselves to blame how Borland lost their hobby market ethos.
What productivity do you gain by using Lispworks? The only benefit is the cross-platofrm graphics api last i checked.
Aside from CAPI, I'd note the excellent platform support (traditionally one would use CCL as a free alternative as SBCL isn't quite there yet) and the possibility to deliver "tiny" executables due to tree shaking.
The class explorer is also excellent for navigating CLOS heavy codebases. Profiler and overall IDE experience is far better for anyone without Emacs Stockholm syndrome (which I do have btw).
The class explorer is also excellent for navigating CLOS heavy codebases. Profiler and overall IDE experience is far better for anyone without Emacs Stockholm syndrome (which I do have btw).
Slightly off the topic: if I have to start using Lisp now, how do I start?
I tried curating things on my own but felt overwhelmed with the alternatives and resources.
Will appreciate your take on this.
Will appreciate your take on this.
Best beginners book http://www.gigamonkeys.com/book/
Curated library list https://awesome-cl.com
Cookbook https://lispcookbook.github.io/cl-cookbook/
IDE+Runtime https://portacle.github.io
Printable cheatsheet http://clqr.boundp.org
Curated library list https://awesome-cl.com
Cookbook https://lispcookbook.github.io/cl-cookbook/
IDE+Runtime https://portacle.github.io
Printable cheatsheet http://clqr.boundp.org
I like Practical Common Lisp if you’re dead set on Common Lisp as your lisp.
However, I’m a fan of How To Design Programs as a gentle introduction, followed by the Racket Guide (EDIT: if you're open to other lisps). Racket is a Scheme-descended lisp, and I feel has better pedagogy, community, and a more modern feel than Common Lisp.
I also like Clojure. There have been several decent books, some free, in addition to Clojure Koans and the 4clojure exercises.
EDIT: While I like Emacs, if you don’t already grok Emacs, DrRacket is a very compelling reason to start your lisp journey with Racket.
However, I’m a fan of How To Design Programs as a gentle introduction, followed by the Racket Guide (EDIT: if you're open to other lisps). Racket is a Scheme-descended lisp, and I feel has better pedagogy, community, and a more modern feel than Common Lisp.
I also like Clojure. There have been several decent books, some free, in addition to Clojure Koans and the 4clojure exercises.
EDIT: While I like Emacs, if you don’t already grok Emacs, DrRacket is a very compelling reason to start your lisp journey with Racket.
If you're starting to learn, please contribute to http://Lisp-Lang.org at the very least by highlighting parts that may need improvement for a new learner by opening an issue at
https://github.com/LispLang/lisplang.github.io
But to answer your question directly, when I was re-learning CL circa 2005 long after university, I populated these pages as I went along:
https://play.org/links/lisp-intro
https://play.org/links/lisp
(I've meant to fold content from those into Lisp-Lang.org but never found/made the time. I did, however, prune obsolete bits earlier this year.)
Best of luck to you!
If you or anyone seeking to learn is in Vancouver, Canada, feel free to reach out and meet for coffee/tea/beer/cider.
But to answer your question directly, when I was re-learning CL circa 2005 long after university, I populated these pages as I went along:
https://play.org/links/lisp-intro
https://play.org/links/lisp
(I've meant to fold content from those into Lisp-Lang.org but never found/made the time. I did, however, prune obsolete bits earlier this year.)
Best of luck to you!
If you or anyone seeking to learn is in Vancouver, Canada, feel free to reach out and meet for coffee/tea/beer/cider.
On Lisp by Paul Graham is a great starting point. It's a quick read, it will give you a conceptual basis, while exciting and inspiring you to use it's power. After that just start using it. As others have mentioned, emacs, slime, etc. are helpful.
I am going through Touretzky's Gentle Introduction to Common Lisp and I recommend it. You can download the entire book in pdf for free, or buy an updated version of the book. I don't see much difference between the two.
The venerable pgloader (https://pgloader.io/) is written/maintained in CL. It's pretty cool.
I'm using Common Lisp for a program I wrote called simple-content-host. I'm connecting Common Lisp and C++17 to create a tiny CDN!
Not only that, but I stream my programming live on Twitch and youtube, check it out:
https://youtu.be/hWmrSyWZS5A
https://www.twitch.tv/kingherring
P.S. here's the code https://github.com/kennymalac/simple-content-host/tree/maste... ;)
P.S. here's the code https://github.com/kennymalac/simple-content-host/tree/maste... ;)
A lot of us at #lispgames[1][2] community using it to make games and related tools. We have a bunch of awesome libraries and wrappers, few game engines/frameworks (3D/2D alike) in development, games too!
Reasons we all chose Common Lisp are often personal, but almost all agree it is the superior interactivity, high language and runtime customizability, performance.
[1]: https://github.com/lispgames/lispgames.github.io/wiki (lispgames.org)
[2]: #lispgames channel at freenode.net
Reasons we all chose Common Lisp are often personal, but almost all agree it is the superior interactivity, high language and runtime customizability, performance.
[1]: https://github.com/lispgames/lispgames.github.io/wiki (lispgames.org)
[2]: #lispgames channel at freenode.net
SISCOG, a Portuguese company, uses Common Lisp on their planning software.
http://www.siscog.eu/
http://www.siscog.eu/upload/GESTAO-DE-LISTAS/News/PDF/eclm-2...
If you feel like enjoying Lisbon or Porto's weather, they are hiring.
http://www.siscog.eu/
http://www.siscog.eu/upload/GESTAO-DE-LISTAS/News/PDF/eclm-2...
If you feel like enjoying Lisbon or Porto's weather, they are hiring.
It's nothing major, but I built a simple web scraper for collecting screenshots from various forums and then creating a static site from it.
I've also used it to build small tools like generating project documentation or collating data from online api's.
Emacs was my gateway (via org mode) and I messed around with Emacs lisp for a while before learning Common Lisp. Developing in Emacs + SLIME is a lot of fun once I get in the rhythm.
I've also used it to build small tools like generating project documentation or collating data from online api's.
Emacs was my gateway (via org mode) and I messed around with Emacs lisp for a while before learning Common Lisp. Developing in Emacs + SLIME is a lot of fun once I get in the rhythm.
As part of a startup, I wrote an automated trading platform in SBCL Common Lisp. It is about 70,000 lines.
I use lisp these days when hacking some automation around common tasks like scraping web sites, extracting info from xlsx or rtf files, and producing LaTeX documentation.
I use lisp these days when hacking some automation around common tasks like scraping web sites, extracting info from xlsx or rtf files, and producing LaTeX documentation.
FYI: there's currently a DNS problem for "sbcl.org", so instead, please use what it ordinarily points to:
https://sbcl.sourceforge.net/
https://sbcl.sourceforge.net/
I use clisp mainly with quicklisp. Lisp isn't rather unique and simple so I have used it for an occasionally algorithm that is annoying any other way. Not for the entire system typically.
Same, for small personal projects only. Drawing a quicklisp map for example:
https://github.com/pablovaldes/quickdot
You can find a picture example of the whole thing in the directory doc
https://github.com/pablovaldes/quickdot
You can find a picture example of the whole thing in the directory doc
Does using StumpWM count?
I think it does! The whole Lisp desktop is a step closer now: StumpWM, Emacs & GuixSD.
It’s kind of funny that those all use divergent versions of Lisp... :)
Doesn’t matter. That’s why it’s called “Common” Lisp. Any code will run on any conformant CL.
Fair point, but in the specific case the differences matter indeed. StumpWM is coded and extensible in Common Lisp, while Emacs use Emacs Lisp and Guix is built on Guile Scheme - both not CL conformant.
it's a very lispy stack, otherwise!
it's a very lispy stack, otherwise!
Ah. Yes. Right. But these aren’t Common Lisps. But you’ve indirectly brought up another useful point, to wit, the general Lisp paradigm is so useful that language designers regularly approximate it when building new platforms. (Clojure is another.)
I've brought or used CL in a few companies.
2005: At a bank dealing with mortgage secondary markets, we snuck it into the Research group (where "research" isn't R&D but feeding information to the trading floor) for plowing through dirty data from multiple databases before our DBAs could make it available via Oracle. This hack helped keep the math & stats PhDs’ pipelines full.
Rationale for using CL: because values have type rather than variables (but may be asserted with CHECK-TYPE; see Google CL style guide), this essentially gave us late-binding of schema.
2006: at Zillow, we were building autonomous server farm control at a time when AWS still emerging. Unfortunately, two things happened at the same time. 1) There was a re-org in Ops leading to discontinuing contractors, so both of us had a 2 weeks to finish it and finish early. 2) SBCL had just reached 1.0 after years in 0.x land. One modification in the run up to SBCL v1.0 that I overlooked in the changelog was how bindings for special vars were handled across threads. It required a trivial fix, but we didn't track that down before time ran out-- oops!
Rationale for using CL: 2 guys still learning the language (one also working on Masters) got lots of traction quickly.
2007: at an education administration start-up, same colleague as from Zillow and I began working on a conceptual sibling to Amazon Dynamo, roughly when their original white paper came out. In 5 weeks, we had an MVP; Kevin wrote all the code, I helped with design and code reviews; it handled arbitrary data payloads and migration of data "ownership" by node. We were about to begin multi-node load & capacity measurements when the business joined a Microsoft incubator and removed all Unix-only folks.
Rationale for using CL: 2 guys, lots of traction quickly... plus, late binding of schema via MOP tricks.
I would like to believe that parts of this continue as Kevin’s vivace-graph despite being a very different animal.
2008: at a Seattle-based ad network (I still bear the shame...), the geek-macho angle was that this is one of the few types of businesses likely to see billions of requests per day. In 4 months, we were deployed at RackSpace and completed baseline load & capacity measurements.
There was an issue with garbage-collection after Hunchentoot would release a HTTP connection, because by the time FreeBSD would release the underlying TCP/IP structs, several minutes would elapse before gc could reclaim everything. Not ideal, but it was manageable!
Rationale for using CL: 1 guy, lots of traction, successful hand-off to new hire (the guy who later brought Clojure to Amazon’s retail side of the house).
Bonus: local Angels and VC (Madrona) came to us but then came the economic crash of 2008...
2009: recruited into Memetrics, a former startup that had been acquired by big consulting firm, and the core software was already written in CL from nearly a decade earlier. I was only in the core group briefly, but subsequent work using CL was my default choice and already approved. (Then again, I didn’t ask permission or beg forgiveness.) There were more dirty data cleaning tools and a Recommender system from which China’s answer to Amazon became a client. Mind you, they didn’t use this code, as they had their own, but it demonstrated that our group understood the principles. I still count this as a win.
Rationale for selecting CL: was decided by folks in Sydney before I arrived. Their head-hunter found me after their acquisition.
2013: Splunk acquired BugSense. BugSense used "Erlang, Lisp and C" (Scheme R4) for handling billions of inbound requests per day non-stop from all time zones into our cluster. While not CL, having previously gone sufficiently deep into CL allowed me getting from zero to presenting at Erlang Factory conference in exactly one year while reaching more than 25x performance increase with the pure-Erlang rewrite.
Rationale for selecting Lisp by their founder: made for malleable query language.
2015: wife wanted to move back to Canada after several years in Silicon Valley. Resurrected a pet project circa 2007, snagz.net, and I attempted founding a company. Back-end is CL for Natural Language Processing work-flow. Core NLP things now use spaCy.io, because their NLP-fu is far better than my self-taught version. Some intermediate working data sets were pushed into a slightly modified Anarki fork of HN, partly to see what the Arc language was about, and partly because that tool fit the (internal) need.
Rationale for selecting CL: Lisp is a beautiful language within which to work.
Others in and around Vancouver using CL: Routific mentions Rust and Common Lisp in their job postings; D-Wave previously mentioned Lisp in job postings but unable to see it today; there was at least one other, but I can’t recall.
In my early experience with Common Lisp, I would have said MACROS were key.
Then, pre-populating memory followed by #’sb-ext:save-lisp-and-die was a handy trick for improving cold-start times. (Alternative to using POSIX mmap().)
Now, it’s simply the joy of writing code in the language.
First milestone: parens disappear, because using the right editor config means never typing a single paren, and editing involves manipulating whole expressions with fluid motion.
Second: thinking how a function should be or might be named, points to something either in the HyperSpec or be the name of a package, thus easy internet search. Today, there’s Quicklisp.org and Quickdocs.org, making this even easier.
That said, I’m currently learning Rust and React-Native for a tiny mobile app, as each language can teach you something new.
But something in the Lisp family will be with me for the rest of my days. I might go with Racket next just to expand my knowledge there.
So I recommend: use Common Lisp for a real project-- just because you can!
2005: At a bank dealing with mortgage secondary markets, we snuck it into the Research group (where "research" isn't R&D but feeding information to the trading floor) for plowing through dirty data from multiple databases before our DBAs could make it available via Oracle. This hack helped keep the math & stats PhDs’ pipelines full.
Rationale for using CL: because values have type rather than variables (but may be asserted with CHECK-TYPE; see Google CL style guide), this essentially gave us late-binding of schema.
2006: at Zillow, we were building autonomous server farm control at a time when AWS still emerging. Unfortunately, two things happened at the same time. 1) There was a re-org in Ops leading to discontinuing contractors, so both of us had a 2 weeks to finish it and finish early. 2) SBCL had just reached 1.0 after years in 0.x land. One modification in the run up to SBCL v1.0 that I overlooked in the changelog was how bindings for special vars were handled across threads. It required a trivial fix, but we didn't track that down before time ran out-- oops!
Rationale for using CL: 2 guys still learning the language (one also working on Masters) got lots of traction quickly.
2007: at an education administration start-up, same colleague as from Zillow and I began working on a conceptual sibling to Amazon Dynamo, roughly when their original white paper came out. In 5 weeks, we had an MVP; Kevin wrote all the code, I helped with design and code reviews; it handled arbitrary data payloads and migration of data "ownership" by node. We were about to begin multi-node load & capacity measurements when the business joined a Microsoft incubator and removed all Unix-only folks.
Rationale for using CL: 2 guys, lots of traction quickly... plus, late binding of schema via MOP tricks.
I would like to believe that parts of this continue as Kevin’s vivace-graph despite being a very different animal.
2008: at a Seattle-based ad network (I still bear the shame...), the geek-macho angle was that this is one of the few types of businesses likely to see billions of requests per day. In 4 months, we were deployed at RackSpace and completed baseline load & capacity measurements.
There was an issue with garbage-collection after Hunchentoot would release a HTTP connection, because by the time FreeBSD would release the underlying TCP/IP structs, several minutes would elapse before gc could reclaim everything. Not ideal, but it was manageable!
Rationale for using CL: 1 guy, lots of traction, successful hand-off to new hire (the guy who later brought Clojure to Amazon’s retail side of the house).
Bonus: local Angels and VC (Madrona) came to us but then came the economic crash of 2008...
2009: recruited into Memetrics, a former startup that had been acquired by big consulting firm, and the core software was already written in CL from nearly a decade earlier. I was only in the core group briefly, but subsequent work using CL was my default choice and already approved. (Then again, I didn’t ask permission or beg forgiveness.) There were more dirty data cleaning tools and a Recommender system from which China’s answer to Amazon became a client. Mind you, they didn’t use this code, as they had their own, but it demonstrated that our group understood the principles. I still count this as a win.
Rationale for selecting CL: was decided by folks in Sydney before I arrived. Their head-hunter found me after their acquisition.
2013: Splunk acquired BugSense. BugSense used "Erlang, Lisp and C" (Scheme R4) for handling billions of inbound requests per day non-stop from all time zones into our cluster. While not CL, having previously gone sufficiently deep into CL allowed me getting from zero to presenting at Erlang Factory conference in exactly one year while reaching more than 25x performance increase with the pure-Erlang rewrite.
Rationale for selecting Lisp by their founder: made for malleable query language.
2015: wife wanted to move back to Canada after several years in Silicon Valley. Resurrected a pet project circa 2007, snagz.net, and I attempted founding a company. Back-end is CL for Natural Language Processing work-flow. Core NLP things now use spaCy.io, because their NLP-fu is far better than my self-taught version. Some intermediate working data sets were pushed into a slightly modified Anarki fork of HN, partly to see what the Arc language was about, and partly because that tool fit the (internal) need.
Rationale for selecting CL: Lisp is a beautiful language within which to work.
Others in and around Vancouver using CL: Routific mentions Rust and Common Lisp in their job postings; D-Wave previously mentioned Lisp in job postings but unable to see it today; there was at least one other, but I can’t recall.
In my early experience with Common Lisp, I would have said MACROS were key.
Then, pre-populating memory followed by #’sb-ext:save-lisp-and-die was a handy trick for improving cold-start times. (Alternative to using POSIX mmap().)
Now, it’s simply the joy of writing code in the language.
First milestone: parens disappear, because using the right editor config means never typing a single paren, and editing involves manipulating whole expressions with fluid motion.
Second: thinking how a function should be or might be named, points to something either in the HyperSpec or be the name of a package, thus easy internet search. Today, there’s Quicklisp.org and Quickdocs.org, making this even easier.
That said, I’m currently learning Rust and React-Native for a tiny mobile app, as each language can teach you something new.
But something in the Lisp family will be with me for the rest of my days. I might go with Racket next just to expand my knowledge there.
So I recommend: use Common Lisp for a real project-- just because you can!
Tankan is a browser-based program for Windows for memorizing Japanese characters.
http://www.kylheku.com/tankan/
It's written in Common Lisp, except for the little tray icon control program that is in Visual C++.
http://www.kylheku.com/tankan/
It's written in Common Lisp, except for the little tray icon control program that is in Visual C++.
I don't know what the relation is between common lisp and scheme but I use R5RS (the S is for "scheme") mainly to program the GIMP to make it do batch image-processing. Gimp also has a Python api, but I just find scheme to be more accessible and user-friendly.