Practical Scheme
Posted2 months agoActive2 months ago
practical-scheme.netTechstoryHigh profile
calmmixed
Debate
60/100
SchemeLispProgramming Languages
Key topics
Scheme
Lisp
Programming Languages
The 'Practical Scheme' website is discussed, with comments ranging from praise for Scheme to criticism of the site's lack of practical examples and concerns about the language's usability and support.
Snapshot generated from the HN discussion
Discussion Activity
Very active discussionFirst comment
1h
Peak period
26
6-12h
Avg / period
8
Comment distribution56 data points
Loading chart...
Based on 56 loaded comments
Key moments
- 01Story posted
Oct 21, 2025 at 1:47 AM EDT
2 months ago
Step 01 - 02First comment
Oct 21, 2025 at 2:49 AM EDT
1h after posting
Step 02 - 03Peak activity
26 comments in 6-12h
Hottest window of the conversation
Step 03 - 04Latest activity
Oct 23, 2025 at 9:50 AM EDT
2 months ago
Step 04
Generating AI Summary...
Analyzing up to 500 comments to identify key contributors and discussion patterns
ID: 45652859Type: storyLast synced: 11/20/2025, 6:36:47 PM
Want the full context?
Jump to the original sources
Read the primary article or dive into the live Hacker News thread when you're ready.
All of the necessary functions to create such a C program documented here: https://cisco.github.io/ChezScheme/csug10.0/foreign.html#./f...
The relevant functions for this applictation are:
Sscheme_init
Sregister_boot_file_bytes
Sbuild_heap
Senable_expeditor (if you want to use Chez's builtin REPL)
Sscheme_start
Sscheme_deinit
I'd post an example program, but I've not got access to my development machine at the moment.
- Bad compilers (that lacked stdlib features)
- Lacked package managers (so we need to detect versions instead of just specifying used libraries for the program)
- The mess of various nix/linux distros having differing paths (/bin or /local/bin or /usr/local/bin or whatever?) and nix binaries lacked a standardized way to just locate "themselves"
In contrast, Windows programs mostly just make an API call to detect where they live and then just load files from relative paths, this also allows for side-by-side installations of varying versions instead of multiple builds (Yes, some programs sadly needed installers but that's just bad engineering, whilst many programs have portable variants).
Considering Mac programs also are self contained I guess those also have some sane API's for program self-location.
Yes, I do realize that much of the centralization of programs harkens back to Unix multi-user paradigms with centralized management but personal computers has been the norm for some almost 40 years at this point (Even if we've moved to web mainframes instead).
I'm guessing it doesn't get much chatter due to INRIA being not very good at promotion of the stuff they do, and Bigloo doesn't have the academia-industry-matrimonial push that e.g. Pharo has received.
https://github.com/babashka/babashka/wiki/Self-contained-exe...
[0]: http://wiki.call-cc.org/man/5/Cross%20development#cross-deve...
I think there's been some work lately to improve Chicken's Windows support.
0 results
why? it's so good
https://practical-scheme.net/wiliki/schemexref.cgi?ChezSchem...
I liked scheme as a learning tool and got highest grades. That doesn't change my impression that most of the sites covering it are mental masturbation. Puzzles, or programs for the sake of scheme itself. Where's the stuff to solve real world problems? (OK, mostly solved by other languages anyway... Still: where is the content that really wants a junior to try it out for routine problems and the senior tonstay with it?)
But for building makets of production systems to see where they are likely to fall over I've never found a better language.
The fact that you have to build everything from scratch is a plus when dealing with the eldrich horrors that lurk in business logic - you think you can use a standard queue? Hahaha let me tell you about a 6pm spike in latency that no one could explain and was driving the cellular network of a tier two city towards failure.
I should probably spell it maquette since a plethora of useless letters is the sine qua non of intellectual sophistication.
This is specifically French linguistic heritage. For a long time the upper classes of England spoke French, and now we're stuck with it. :)
Hwail ai kenker wiq yer kr1tes1zm ev i6g1c orxogrefi, w1tc servz lardjli ez e wei te s1gnel socel kl4s bai wei ev i1rz weisted m3meraizi6 iuslesli 1r3giuler sp3li6z, iu k4nt boil q4t ocen 1n 3vri kament. Bet iu mei bi 1nter3sted 1n http://canonical.org/~kragen/alphanumerenglish. 1t simz laik qe kaind ev xi6 m3ni skimerz w5d laik, r1li.
It's easy to teach any programmer Scheme sufficient for maintenance. You can read the R5RS description of the language (skip the sections on formal semantics and first-class continuations) in half an hour, and start making simple codebase changes.
Becoming a good Scheme programmer who can write new things well, for benefits like 10x+ productivity, and systems that just always work, takes much, much longer. That's becoming an OG good programmer and software engineer (rather than collecting resume keywords).
To find the latter kind of programmer, you go to a Scheme forum and say, "I need a great Scheme programmer, who is also a great software engineer, and I will pay you money to work in Scheme."
I'm not sure what you mean by "real world problems" but I think most would consider Cisco router firmware to be in that domain. In some sense, due to Cisco the Internet runs on two Lisp-like programming languages: Erlang and Chez.
If you look at TFA you'll find that it links over to a few articles that describe inventing a Scheme to solve things like high paced computer graphics production and large scale inventories.
I'm genuinely confused why you would be convinced about this. What are the points of similarity that aren't common to all functional languages?
What I meant was: where are the resources that teach how to tackle everyday chores? O'Reilly has a lot of "Real world <niche lang.>".
Not finished "practical" software -- albeit it's utterly cool to see that there are working projects in numbers / good showcases.
It's fine to have a collection for scheme (like the endless and sometimes helpful "awesome x" collections).
I'm missing the "Automate the boring stuff" and the like.
Maybe I'm more irritated about the lack of adoption (and grumpy about that -- not really the OP). E.g. I don't get it that Nix has more outreach than Guix, despite even Nix-users sometimes agree that the language isn't a strong selling point (I don't know about the idiosyncrasies of Guile, seemed preferable at first glance).
The nice thing about the Scheme community's aggressive inability to do practice advocacy is that there's very little noise like you get in employable languages.
For example, if you Google something about Scheme, there won't be a thousand redundant SEO 'tutorials' that were written in bad faith to attract eyeballs, rather than to fill a need and inform. (The closest Scheme comes to noise is when bloggers get a blog post out of trying Scheme, but most of the rest of it tends to be high quality relative to popular languages.)
> One day, however, I will point this page, when the friend asks me if Scheme is feasible for daily chores and a practical choice.
Atheists in 1999 might have to go Usenet comp.lang.scheme to find Scheme experts.
(Scheme polo shirt at church in 1999? My first guess is around Rice University. Second guess is Indiana.)
And of course MIT. https://archivesspace.mit.edu/repositories/2/archival_object...
I was at MIT in 1999, and class T-shirts occasionally happened, and there were tens of churches within walking distance, for historic reasons, but not a lot of churchgoing people from the universities themselves, that I'm aware of.
But who knows: toplevel works in strange and mysterious ways.
> My work place was using it in production, 1999. A ton of code was written by a VERY SMART (and famous) person and of course it worked. He delivered it under pressure, ahead of schedule and it just worked. Ok, but my frustration was that we could not find anyone to support the decoders.
Actually, except for the "famous" part, that sounds a bit like a major system in Scheme for which I was hired by a very smart person who'd done 10x or 100x development on it. And it just worked, and we could evolve it rapidly, and keep it just working. In 1999, it probably would have been based in New Jersey.