I had seen Lobster before, but not really looked closely. Seeing it again now, I think I was wrong to dismiss it. Just at the syntactic level with semantics described in the link, it looks like it really might be "Python done right". The link mentions lots of features, but the following bits caught my eye.
The let/var declarations for constants/variables is much better than implicit declaration, which silently hides typos and necessitates ugly global/nonlocal declarations. (Mojo offers this improvement too.)
I don't know for sure, but it seems like it's embraced block arguments comparable to how Ruby or SmallTalk does it. So you can add your own control flow, container visitors, etc. I think of this as another syntax for passing a lambda function as an argument, and I'm curious if Lobster's optimizer flattens it to a basic block when possible.
I think I'll try to learn more about it. I wonder if the name is a nod to Accelerando.
Looks like a cool language. Like the short syntax. Impressed they implemented so much of OpenGL natively in the language. Lot of work to implement OpenGL.
However, the real test, from personal perspective, especially with a custom framework for shaders, would be implementing one of the medium difficulty LearnOpenGL 3D examples, such as the Multiple Lights example. https://learnopengl.com/Lighting/Multiple-lights
With how much goes wrong with OpenGL and shader development, wary of delving very much into an OpenGL centric language, and then finding out there's a bunch of gotchas in the part that tends to be desirable (3D rendering). "Whoops, one of a zillion hidden flags doesn't get set properly for some reason."
Texture loading seems like its there, and it says it actually uses "stb_image.h" internally. Plus, it apparently does cubemaps.
Lights seem implemented, although not sure what that does "sets up a light at the given position for this frame."
Model Loading is also another, from the built-ins it looks like it does .ply and .iqm? (Inter-Quake Model) files.
Matrice math has a few although seems like a LookAt and Perspective matrix creator in the 4x4 category would be needed that return a matrix you can apply to a shader uniform. gl.perspective looks weird, just says "changes from 2D mode (default) to 3D right handed perspective mode"
Either way, looks cool, would just like to have more examples in the target demographic, OpenGL game development. Even if they're simple. That's part of why the LearnOpenGL examples are so useful. Simple. Implementing even a significant subset would go along way toward selling me on game development in an unknown language.
the "gl" namespace is really a bit of misnomer since it is much higher level API than OpenGL, and does not really depend on OpenGL. It could pretty easily run on another API underneath.
The repo comes with a LOT of graphical samples (in the "samples" dir).
The lights are more of standard way of passing them to the shader (see the phong shader implementation).
Matrices are a bit of a weak spot since its all implicit (which is convenient, but not powerful). There are also classes for direct use of matrices though.
Cool. Thanks for the reply and clarification. Although the clarification is admittedly confusing when it's called the "gl" namespace.
Not seeing a phong example in the samples dir. "pendulum.lobster", "physics_boxes.lobster" Only reference to "phong" brought back in a search was milen-prg's comment on Aug 7, 2023 about using the phong shader in the "cgtest.lobster" example.
Well, it has an animal mascot logo. Which is my personal yardstick for if a project is destined for success. So, off to a good start, but the lobster could be more cuddly.
Fun story behind the name, per the language inventor Hendrik K. Grubbs, Phd.
"Dr. Grubbs, what inspired the language name"
"A Hooker down by the docks"
"WHAT?!?"
"Yes, I went to her lonely one Friday, she only cost five dollars. Next day I had a terrible itch and realized she'd given me crabs. So I confronted her the next day and asked why she had done this to me. 'What do you expect for five dollars, lobster?' "
To be fair, there’s a programming language out there already with a cuddly-lobster mascot. A decidedly un-cuddly lobster just maximizes product differentiation.
This is a really nice looking language. Feedback in case the creator sees but it wasn't obvious to me at first that it was targeting game development. The first mention is in features:
> Features have been picked for their suitability in a game programming language
Would be fun to see some basic games like tetris, pong etc in Lobster in case anyone has an example floating round?
It comes with a tiny minecraft clone in < 100 lines of Lobster.
Agreed it could use more actual game examples. I've written a ton of game prototypes in it, and some could do with open sourcing, just haven't gotten around to it.
I noticed that Lobster, like many other languages, does not use the C style for-loop syntax, `for (int i = 0; i < m; i++): print(i);`. Could you say why that is? I have been writing Python for many years but still miss that C syntax. Initiating `i` right before every loop feels clunky to me.
In Lobster, this is `for(m) i: print(i)`
Or simply `for(m): print(_)`
The syntax comes from the observation that even in C-like languages, 99% of the time I want to iterate over the range (0..m-1), so that's what it is optimized for.
I don't enjoy writing the same boilerplate over and over. Moreover, because it is so verbose, it is hard to spot `for (int i = 0; i <= m; i++): print(i);` being a special kind of loop since it looks so similar to the common `<` case. I'd rather that looks entirely different to alert me we're also iterating over the bound.
I am not sure what you mean by "Initiating `i` right before every loop"
Obviously some will find this a silly opinion but the one thing that turned me off the most about the Nim programming language was its use of significant whitespace. The same is true with F# (and of course Python). Having had apps with YAML for config, and having had nightmares trying to copy/paste config directives from various sources, I just find whitespace to be unwieldy.
Now that's a strong opinion, (weakly held - as a language can't be judged based on this design decision). But it does sour my interest a bit.
> Having had apps with YAML for config, and having had nightmares trying to copy/paste config directives from various sources, I just find whitespace to be unwieldy.
Convert your YaML into JSON and save it in your YaML file. There is probably an online converter, but writing one in your language of choice should be less than ten lines of code.
Do the same YaML→JSON for the “source” configuration you want to copy from, and copy-paste the parts you want. Leave them as JSON.
Complaining about Python's significant whitespace, I get it. I don't mind it personally, but it's obligatory and you can't overcome it (unless you do `coding: with_braces` tricks, of course). But why one would complain about YaML's whitespace? It is not obligatory.
> But why one would complain about YaML's whitespace? It is not obligatory.
The problem (as felt by me and also as identified by the person you replied to) is that you can't copy-paste/munge some stuff into the right spot and then just let the formatter to fix the indentation. It's not a problem that the format "at rest" has whatever certain indentation to be correct, its that while being actively editing your formatter cannot automatically set the correct indentation.
The flow that you're talking about of converting yaml to json and then putting it into yaml could work in some cases but thats very much a kludge. It will have numerous bad side effects unavoidable, including that it would discard comments in the middle since JSON doesn't allow for comments at all, theres no timestamps in JSON, there's no octal numbers, etc.
> The problem (as felt by me and also as identified by the person you replied to) is that you can't copy-paste/munge some stuff into the right spot and then just let the formatter to fix the indentation.
That problem I undestand, and that is why I suggested to convert both into JSON —or YaML with default_flow_style=True which would preserve datetimes and other non-JSON stuff— and copy-paste without the hassle of having to indent/unindent correctly. Of course that doesn't help with copying comments. That would need extra copy-paste operations, but still one hasn't the hassle of significant whitespace. The following is also valid YaML:
{"some_key": {
"attr1":
# an intermittent comment
"val1", "attr2": 12312 # more comments!
}
}
My gripe with json is the lack of support for comments. Whenever I come across a config file that has comments about what the config line(s) mean, I am so grateful.
Whenever I come across a json config file, I kind of despair a little and start poking at the code in hopes there are comments about what the config means.
I totally agree with your gripe about JSON's lack of comments. There were people AFAIK who tried to write a spec with comments (and maybe dangling commas? was it called JSON5?) but by then it probably was too late.
My biggest issue with JSON5 is as far as I'm aware, if you update settings programmatically, you tend to lose comments... not sure of any implementation that preserves them.
Never understood how putting up roadblocks for developers trying to copy-paste code was deemed acceptable, or GVR (and others) thought the solution to poorly formatted code was making formatting carry semantics instead of just writing an auto formatter.
Well of course that's completely ahistorical. The motivator was eliminating block terminators. Python's predecessor, ABC, was a teaching language; programs were small and refactoring was not a thing. Heck, copy-paste was hardly a thing, and there certainly were no auto formatters. Indented code was a step up from BASIC, the teaching language prevalent at the time.
I agree. Autoformatters are everywhere and easy to use. I'd far rather do that (plus maybe a pre-commit hook) than have to deal with whitespace in the language.
Yea, in the 90s significant whitespace seemed great because it meant that you got readable code. The amount of code that you might see copy/pasted with terrible formatting/indentation in other languages could make you want to scream.
Now, when you paste code and things are wrong, an auto formatter cleans it up for you. Before, you'd just end up with an unreadable codebase.
If you grab that version and unpack it and look at /OChangelog then it seems to date back until at least 1989, same as Python itself.
That was for C source, of course. I expect there were pre-GNU indent variants, perhaps posted on comp.sources.unix and maybe some commercial things as part of very expensive compiler packages.
I would say that running autoformatters in any kind of routine way was pretty rare. EDIT: but I think ascribing the language design to commonality or not is probably ahistorical. Even today it's a rather passionate debate. And even at the time, Lisp - the poster child of copy-paste friendly PLangs - was routinely autoformatted within Emacs', but that was not enough for people to not find Lisp code "ugly".
I get everyone has their thing, but I've been writing Python professionally for years and I can't even remember the last time significant white space was an issue. You just get used to it, like everything else.
I'm automatically going to be interested in any language with significant white space because there are very few mainstream ones and I hate the visual clutter that block delimiters create. Pretty much there's just Python. Scala 3 can happily do both.
I think we'd be better off if text editors just had option of representing braces and such as consistent indentation. Block delimiting tokens should optionally have semantics of non directly printable characters like new line or tab.
You'd love Haskell, which uses curly braces for many constructs, but also has rules by which they are implied by indentation -- so in practice you only ever see them on records.
I love python syntax overall, absolutely despise Haskell. Wastes my time constantly and gives me incomprehensible compiler errors when you screw it up. Expression oriented languages are really poorly suited for whitespace imo, unless they're hyper-regular like s-expressions: I could imagine a decent whitespace-based version of those.
A language that can do both Python and C "styles" is Ring. It is possible. But the issue is people have such a strong preference for one or the other, that they force the language and developers to permanently choose.
Even Allman versus K&R or tabs versus spaces are huge battles, without even going into significant white space.
Yes, its somewhat similar, monomorphic specialization.
I am sure it has some limitations, but it is pretty powerful, it can even do things C++ can't, like have a type error in a branch that in the current specialization is never executed doesn't count as a type error :)
There's no FPL style pattern matching. It can do a switch case on types (the various subclasses of a base class) but only 1 level, i.e. can't match against type of members. That could still be added, just personally rarely have use for it.
> like have a type error in a branch that in the current specialization is never executed doesn't count as a type error
I’ve struggled with my feelings on this even to the extent that C++ allows it, because while it is flexible, it can also hide errors in libraries that will only blow up when used in very specific ways.
Yes, that is a downside, but when you write very generic code being limited in what code you can write because it has to typecheck even for types that it doesn't apply to is one of the more frustrating things of doing this in C++ and other languages.
I agree with your point in general, and I'm curious if it's a problem in Lobster. Here's one in Rust:
let i = 1;
let j = 1;
print!("i: {:?}\n", !i);
print!("j: {:?}\n", !j);
// pretend there's a lot of code here
// spooky action at a distance
let v = vec![1, 2, 3];
v[i];
You might think those two print statements would print the same value. They do not.
In Lobster you must initialize a variable declaration. Now you can produce your bad case with `var x = nil`, but that is undesirable because nil types must be explicitly checked at compile time, so typically you want to use as few nils as possible. More likely thus is `var x = ""` if you must initialize later.
For case like this, I'd say your text editor should definitely just be able to tell you right away that this variable is a "string" when you mouse over it.
One thing I hate about Generative AI is that it has flipped the value prop of making your language similar to an existing popular language. This helps new programmers but it really messes with generative AI. I can feel the era of fun new programming languages that might break big ending.
I'm afraid about that too, but I hope that AI will get significantly smarter faster than it takes for any new language take popularity among humans. That it will be smart enough to not be so language sensitive.
Who knows. Maybe some new fun language will pop up that's hard to write for humans, but easy to write for AI (because it can work in millisecond loop with language server, think borrow checker to the moon) and also exceedingly easy to read for humans. Because humans will, I think stil for a long time, need to debug ever shrinking corner cases where AI generated something subtly but spectacularly wrong.
I don't know. I wrote a compiler with a syntax that's close to all the curly languages, but different in some ways. ChatGPT had no problem cranking out some test cases after I explained the rules and gave minimal examples. I guarantee it wasn't trained on any source code of my language, because none existed more than two months ago.
Running some of the samples feels like those old SGI Iris GL demos, but implemented with a way nicer and concise language. I'll try to port some of my old code from that era.
This feels like a scripting companion to Rust with Python-like syntax, nice one. I definitely could see it being using as a Lua replacement in embedded contexts.
I had misunderstood a comment elsewhere and had claimed that you were the author of Quake (which I now know was developed by a team). My link above acknowledges that you are the author of some Quake maps (among many other things).
I had seen Lobster before, but not really looked closely. Seeing it again now, I think I was wrong to dismiss it. Just at the syntactic level with semantics described in the link, it looks like it really might be "Python done right". The link mentions lots of features, but the following bits caught my eye.
The let/var declarations for constants/variables is much better than implicit declaration, which silently hides typos and necessitates ugly global/nonlocal declarations. (Mojo offers this improvement too.)
I don't know for sure, but it seems like it's embraced block arguments comparable to how Ruby or SmallTalk does it. So you can add your own control flow, container visitors, etc. I think of this as another syntax for passing a lambda function as an argument, and I'm curious if Lobster's optimizer flattens it to a basic block when possible.
I think I'll try to learn more about it. I wonder if the name is a nod to Accelerando.
Yup, it generally inlines lambdas passed to custom control flow functions, so it ends up pretty similar to hand-written.
The name does not refer to anything in particular :)
Related. Others?
The Lobster Programming Language - https://news.ycombinator.com/item?id=44051841 - May 2025 (6 comments)
The Lobster Programming Language - https://news.ycombinator.com/item?id=31453822 - May 2022 (14 comments)
The Lobster Programming Language - https://news.ycombinator.com/item?id=25498005 - Dec 2020 (4 comments)
The Lobster Programming Language - https://news.ycombinator.com/item?id=19567160 - April 2019 (164 comments)
The Lobster Programming Language - https://news.ycombinator.com/item?id=15557060 - Oct 2017 (2 comments)
Looks like a cool language. Like the short syntax. Impressed they implemented so much of OpenGL natively in the language. Lot of work to implement OpenGL.
With such a focus, be nice to have a few more OpenGL examples. Took a bit of looking, yet found a longer example for a 2D shooter. https://aardappel.github.io/lobster/shooter_tutorial.html
However, the real test, from personal perspective, especially with a custom framework for shaders, would be implementing one of the medium difficulty LearnOpenGL 3D examples, such as the Multiple Lights example. https://learnopengl.com/Lighting/Multiple-lights
With how much goes wrong with OpenGL and shader development, wary of delving very much into an OpenGL centric language, and then finding out there's a bunch of gotchas in the part that tends to be desirable (3D rendering). "Whoops, one of a zillion hidden flags doesn't get set properly for some reason."
Texture loading seems like its there, and it says it actually uses "stb_image.h" internally. Plus, it apparently does cubemaps.
Lights seem implemented, although not sure what that does "sets up a light at the given position for this frame."
Model Loading is also another, from the built-ins it looks like it does .ply and .iqm? (Inter-Quake Model) files.
Matrice math has a few although seems like a LookAt and Perspective matrix creator in the 4x4 category would be needed that return a matrix you can apply to a shader uniform. gl.perspective looks weird, just says "changes from 2D mode (default) to 3D right handed perspective mode"
Either way, looks cool, would just like to have more examples in the target demographic, OpenGL game development. Even if they're simple. That's part of why the LearnOpenGL examples are so useful. Simple. Implementing even a significant subset would go along way toward selling me on game development in an unknown language.
the "gl" namespace is really a bit of misnomer since it is much higher level API than OpenGL, and does not really depend on OpenGL. It could pretty easily run on another API underneath.
The repo comes with a LOT of graphical samples (in the "samples" dir).
The lights are more of standard way of passing them to the shader (see the phong shader implementation).
Matrices are a bit of a weak spot since its all implicit (which is convenient, but not powerful). There are also classes for direct use of matrices though.
Cool. Thanks for the reply and clarification. Although the clarification is admittedly confusing when it's called the "gl" namespace.
Not seeing a phong example in the samples dir. "pendulum.lobster", "physics_boxes.lobster" Only reference to "phong" brought back in a search was milen-prg's comment on Aug 7, 2023 about using the phong shader in the "cgtest.lobster" example.
if you find in files for "phong" you should be able to find samples that use it.. the actual shader is in data/shaders/default.materials
Do you intend on adding a phys namespace, for stuff like jolt or bullet?
Would love to add Jolt at some point, yes.
There's already bindings for Box2D in the "ph" namespace.
Well, it has an animal mascot logo. Which is my personal yardstick for if a project is destined for success. So, off to a good start, but the lobster could be more cuddly.
Can I interest you in a rat named Keith, with its foot blown off?
At least it has personality. Not a generic, noncommittal blob.
Fun story behind the name, per the language inventor Hendrik K. Grubbs, Phd.
"Dr. Grubbs, what inspired the language name"
"A Hooker down by the docks"
"WHAT?!?"
"Yes, I went to her lonely one Friday, she only cost five dollars. Next day I had a terrible itch and realized she'd given me crabs. So I confronted her the next day and asked why she had done this to me. 'What do you expect for five dollars, lobster?' "
To be fair, there’s a programming language out there already with a cuddly-lobster mascot. A decidedly un-cuddly lobster just maximizes product differentiation.
It's a crab, not a lobster.
What language is that?
They mean Rust, but its unofficial mascot is a cuddly crab, not a lobster (both are crustaceans).
This is a really nice looking language. Feedback in case the creator sees but it wasn't obvious to me at first that it was targeting game development. The first mention is in features:
> Features have been picked for their suitability in a game programming language
Would be fun to see some basic games like tetris, pong etc in Lobster in case anyone has an example floating round?
Seen by creator :)
It comes with a tiny minecraft clone in < 100 lines of Lobster.
Agreed it could use more actual game examples. I've written a ton of game prototypes in it, and some could do with open sourcing, just haven't gotten around to it.
Hi, very nice language!
I noticed that Lobster, like many other languages, does not use the C style for-loop syntax, `for (int i = 0; i < m; i++): print(i);`. Could you say why that is? I have been writing Python for many years but still miss that C syntax. Initiating `i` right before every loop feels clunky to me.
In Lobster, this is `for(m) i: print(i)` Or simply `for(m): print(_)`
The syntax comes from the observation that even in C-like languages, 99% of the time I want to iterate over the range (0..m-1), so that's what it is optimized for.
I don't enjoy writing the same boilerplate over and over. Moreover, because it is so verbose, it is hard to spot `for (int i = 0; i <= m; i++): print(i);` being a special kind of loop since it looks so similar to the common `<` case. I'd rather that looks entirely different to alert me we're also iterating over the bound.
I am not sure what you mean by "Initiating `i` right before every loop"
Ah hey, nice language! A minecraft clone in <100 lines sounds like a pretty awesome feat, I'll definitely check it out!
https://github.com/aardappel/lobster/blob/master/samples/lob...
It brings the joy back to graphics and game programming.
Obviously some will find this a silly opinion but the one thing that turned me off the most about the Nim programming language was its use of significant whitespace. The same is true with F# (and of course Python). Having had apps with YAML for config, and having had nightmares trying to copy/paste config directives from various sources, I just find whitespace to be unwieldy.
Now that's a strong opinion, (weakly held - as a language can't be judged based on this design decision). But it does sour my interest a bit.
> Having had apps with YAML for config, and having had nightmares trying to copy/paste config directives from various sources, I just find whitespace to be unwieldy.
Convert your YaML into JSON and save it in your YaML file. There is probably an online converter, but writing one in your language of choice should be less than ten lines of code.
Do the same YaML→JSON for the “source” configuration you want to copy from, and copy-paste the parts you want. Leave them as JSON.
Complaining about Python's significant whitespace, I get it. I don't mind it personally, but it's obligatory and you can't overcome it (unless you do `coding: with_braces` tricks, of course). But why one would complain about YaML's whitespace? It is not obligatory.
is equivalent to is equivalent to is equivalent to and they're all valid YaML (and on the plus side you can leave dangling commas at the end of sequences, but it won't be valid JSON anymore).> But why one would complain about YaML's whitespace? It is not obligatory.
The problem (as felt by me and also as identified by the person you replied to) is that you can't copy-paste/munge some stuff into the right spot and then just let the formatter to fix the indentation. It's not a problem that the format "at rest" has whatever certain indentation to be correct, its that while being actively editing your formatter cannot automatically set the correct indentation.
The flow that you're talking about of converting yaml to json and then putting it into yaml could work in some cases but thats very much a kludge. It will have numerous bad side effects unavoidable, including that it would discard comments in the middle since JSON doesn't allow for comments at all, theres no timestamps in JSON, there's no octal numbers, etc.
> The problem (as felt by me and also as identified by the person you replied to) is that you can't copy-paste/munge some stuff into the right spot and then just let the formatter to fix the indentation.
That problem I undestand, and that is why I suggested to convert both into JSON —or YaML with default_flow_style=True which would preserve datetimes and other non-JSON stuff— and copy-paste without the hassle of having to indent/unindent correctly. Of course that doesn't help with copying comments. That would need extra copy-paste operations, but still one hasn't the hassle of significant whitespace. The following is also valid YaML:
My gripe with json is the lack of support for comments. Whenever I come across a config file that has comments about what the config line(s) mean, I am so grateful.
Whenever I come across a json config file, I kind of despair a little and start poking at the code in hopes there are comments about what the config means.
I totally agree with your gripe about JSON's lack of comments. There were people AFAIK who tried to write a spec with comments (and maybe dangling commas? was it called JSON5?) but by then it probably was too late.
My biggest issue with JSON5 is as far as I'm aware, if you update settings programmatically, you tend to lose comments... not sure of any implementation that preserves them.
Never understood how putting up roadblocks for developers trying to copy-paste code was deemed acceptable, or GVR (and others) thought the solution to poorly formatted code was making formatting carry semantics instead of just writing an auto formatter.
Well of course that's completely ahistorical. The motivator was eliminating block terminators. Python's predecessor, ABC, was a teaching language; programs were small and refactoring was not a thing. Heck, copy-paste was hardly a thing, and there certainly were no auto formatters. Indented code was a step up from BASIC, the teaching language prevalent at the time.
I agree. Autoformatters are everywhere and easy to use. I'd far rather do that (plus maybe a pre-commit hook) than have to deal with whitespace in the language.
To be fair to GvR, autoformatters weren’t commonplace in the late 80s and early 90s. Were there even any?
Ever since Go got big, though, everyone else is discovering how fantastically nice they are, and that’s a good thing.
Yea, in the 90s significant whitespace seemed great because it meant that you got readable code. The amount of code that you might see copy/pasted with terrible formatting/indentation in other languages could make you want to scream.
Now, when you paste code and things are wrong, an auto formatter cleans it up for you. Before, you'd just end up with an unreadable codebase.
It's definitely an odd choice to make now.
GNU indent was already at version 1.9.1 by 1994: https://ftp.gnu.org/gnu/indent/
If you grab that version and unpack it and look at /OChangelog then it seems to date back until at least 1989, same as Python itself.
That was for C source, of course. I expect there were pre-GNU indent variants, perhaps posted on comp.sources.unix and maybe some commercial things as part of very expensive compiler packages.
I would say that running autoformatters in any kind of routine way was pretty rare. EDIT: but I think ascribing the language design to commonality or not is probably ahistorical. Even today it's a rather passionate debate. And even at the time, Lisp - the poster child of copy-paste friendly PLangs - was routinely autoformatted within Emacs', but that was not enough for people to not find Lisp code "ugly".
While formatters for non-C languages may have existed, no auto formatters existed. And yes, this discussion is completely ahistorical.
I get everyone has their thing, but I've been writing Python professionally for years and I can't even remember the last time significant white space was an issue. You just get used to it, like everything else.
I'm automatically going to be interested in any language with significant white space because there are very few mainstream ones and I hate the visual clutter that block delimiters create. Pretty much there's just Python. Scala 3 can happily do both.
I think we'd be better off if text editors just had option of representing braces and such as consistent indentation. Block delimiting tokens should optionally have semantics of non directly printable characters like new line or tab.
You'd love Haskell, which uses curly braces for many constructs, but also has rules by which they are implied by indentation -- so in practice you only ever see them on records.
I'm not sure. The semantics is too wild to care about indentation or delimiters. I love Scala 3 though. Very rich and flexible language.
I love python syntax overall, absolutely despise Haskell. Wastes my time constantly and gives me incomprehensible compiler errors when you screw it up. Expression oriented languages are really poorly suited for whitespace imo, unless they're hyper-regular like s-expressions: I could imagine a decent whitespace-based version of those.
A language that can do both Python and C "styles" is Ring. It is possible. But the issue is people have such a strong preference for one or the other, that they force the language and developers to permanently choose.
Even Allman versus K&R or tabs versus spaces are huge battles, without even going into significant white space.
YAML has given me eye-twitching ever since I went on an ill-considered quest of setting up wifi on a Debian server years ago.
I never figured it out by the way - just bought a really long LAN cable.
YAML with Go templating (like you'd find in Helm Charts) was enough to push me over the edge.
Was there any reason not to use flow collection style, which would free the templates from significant whitespace?
Was coming here to comment the exactly same thing. Significant indentation makes me shudder.
+1 white space significance brings back whole tab vs space preferences and make it much harder for automatic re-formatting as well
Nice! So it looks like polymorphism is done via C++ template-style ad-hoc polymorphism? Are there any restrictions on it?
Also, is there any kind of sophisticated pattern matching? I feel like for me a language without pattern matching is a non-starter these days.
Yes, its somewhat similar, monomorphic specialization.
I am sure it has some limitations, but it is pretty powerful, it can even do things C++ can't, like have a type error in a branch that in the current specialization is never executed doesn't count as a type error :)
There's no FPL style pattern matching. It can do a switch case on types (the various subclasses of a base class) but only 1 level, i.e. can't match against type of members. That could still be added, just personally rarely have use for it.
> like have a type error in a branch that in the current specialization is never executed doesn't count as a type error
I’ve struggled with my feelings on this even to the extent that C++ allows it, because while it is flexible, it can also hide errors in libraries that will only blow up when used in very specific ways.
Yes, that is a downside, but when you write very generic code being limited in what code you can write because it has to typecheck even for types that it doesn't apply to is one of the more frustrating things of doing this in C++ and other languages.
Feels like it’s taking the best of Rust and Ruby (with Python style whitespace)
Admittedly it’s just a first impression
> Flow-Sensitive Type-Inference
imho, I don't consider Type-inference as a good thing when it happens from 50 lines ahead/below. How would regular people follow along?
Good case
x = "hello" // infer type as string - good thing.
Bad case
var/declare x;
50 lines later
if (....)
I agree with your point in general, and I'm curious if it's a problem in Lobster. Here's one in Rust:
You might think those two print statements would print the same value. They do not.In Lobster you must initialize a variable declaration. Now you can produce your bad case with `var x = nil`, but that is undesirable because nil types must be explicitly checked at compile time, so typically you want to use as few nils as possible. More likely thus is `var x = ""` if you must initialize later.
For case like this, I'd say your text editor should definitely just be able to tell you right away that this variable is a "string" when you mouse over it.
It shouldn't require a fancy forward-lookup-capable editor / language-server to show type. That is the point I am trying to make
var/declare x;
25 lines later
call f(x); // ** Reader has no idea what x is ... even though compiler has **
25 lines later
if (....)
One thing I hate about Generative AI is that it has flipped the value prop of making your language similar to an existing popular language. This helps new programmers but it really messes with generative AI. I can feel the era of fun new programming languages that might break big ending.
I'm afraid about that too, but I hope that AI will get significantly smarter faster than it takes for any new language take popularity among humans. That it will be smart enough to not be so language sensitive.
Who knows. Maybe some new fun language will pop up that's hard to write for humans, but easy to write for AI (because it can work in millisecond loop with language server, think borrow checker to the moon) and also exceedingly easy to read for humans. Because humans will, I think stil for a long time, need to debug ever shrinking corner cases where AI generated something subtly but spectacularly wrong.
I don't know. I wrote a compiler with a syntax that's close to all the curly languages, but different in some ways. ChatGPT had no problem cranking out some test cases after I explained the rules and gave minimal examples. I guarantee it wasn't trained on any source code of my language, because none existed more than two months ago.
Running some of the samples feels like those old SGI Iris GL demos, but implemented with a way nicer and concise language. I'll try to port some of my old code from that era.
This feels like a scripting companion to Rust with Python-like syntax, nice one. I definitely could see it being using as a Lua replacement in embedded contexts.
You should take a look at Koto Programming Language, I’ve been trying to keep an eye on it.
https://koto.dev/
One of the few light excellent C++ projects
.
Surely you don't mean Quake the 1996 video game, so what is "Quake" in this context?
I was mistaken ... not the author: https://quake.fandom.com/wiki/Wouter_Van_Oortmerssen_(Aardap...
Yes the author :) I make both programming languages and games / engines / maps..
I had misunderstood a comment elsewhere and had claimed that you were the author of Quake (which I now know was developed by a team). My link above acknowledges that you are the author of some Quake maps (among many other things).