* A "fmt" (presumably "format") package is not where I would expect to find a send-to-stdout function. I'd expect that in "io", and I'd expect "fmt" to provide helpers for actually creating the string to output. (Which would be that much more necessary here, since I assume you don't intend to emulate anything like all the bells and whistles of Python's built-in `print`)
* I don't know where `stringify` is supposed to come from. But the Pythonic way is that you just call the `str` type/constructor. That seems at least as doable as a `stringify` function; either way you're presumably stuck with static overloads for built-in types. (It doesn't look like you plan on supporting a protocol for user-defined type conversions?)
* Putting a string together like this is really unpleasant. I'd advise looking into the new template strings in Python (https://peps.python.org/pep-0750/); they form a solid basis for all kinds of other formatting. The important work is done at compile time; it generates code to create an object using current variable values, packing them for later formatting. For type-safety reasons I suppose the interpolated values have to be coerced to string at compile-time as well, but storing the values this way allows for choosing a different algorithm for assembling the final string (e.g., one that does quoting and escaping for some particular environment).
I could imagine having something like
use otter:fmt
use otter:io
# ...
io.println(fmt.format(t"Point: ({p.x}, {p.y}), distance: {dist}"))
where the t-string gets converted at compile time to something like
(Empty strings are preserved in the sequence so that formatting code knows what was literal and what came from an interpolated value.)
Of course, if you have other approaches in mind for things like type-safe SQL query generation, this could be simplified by just producing the string concatenation directly and avoiding the need for a separate formatter function etc.
Made just for fun or any issues with other languages that it tries to solve? You say "scripting language that compiles". That mean it has fast compilation and meant to use a "$lang run" shebang similar to Nim/Go/etc?
Also I went through SPy's repo, and also looked at what they beat us in, we have advantages in certain areas. But so do they, so i've already planned new features to make our language even more powerful!
> Python-like readability and...
Some thoughts on the syntax, naming etc.:
* A "fmt" (presumably "format") package is not where I would expect to find a send-to-stdout function. I'd expect that in "io", and I'd expect "fmt" to provide helpers for actually creating the string to output. (Which would be that much more necessary here, since I assume you don't intend to emulate anything like all the bells and whistles of Python's built-in `print`)* I don't know where `stringify` is supposed to come from. But the Pythonic way is that you just call the `str` type/constructor. That seems at least as doable as a `stringify` function; either way you're presumably stuck with static overloads for built-in types. (It doesn't look like you plan on supporting a protocol for user-defined type conversions?)
* Putting a string together like this is really unpleasant. I'd advise looking into the new template strings in Python (https://peps.python.org/pep-0750/); they form a solid basis for all kinds of other formatting. The important work is done at compile time; it generates code to create an object using current variable values, packing them for later formatting. For type-safety reasons I suppose the interpolated values have to be coerced to string at compile-time as well, but storing the values this way allows for choosing a different algorithm for assembling the final string (e.g., one that does quoting and escaping for some particular environment).
I could imagine having something like
where the t-string gets converted at compile time to something like (Empty strings are preserved in the sequence so that formatting code knows what was literal and what came from an interpolated value.)Of course, if you have other approaches in mind for things like type-safe SQL query generation, this could be simplified by just producing the string concatenation directly and avoiding the need for a separate formatter function etc.
Made just for fun or any issues with other languages that it tries to solve? You say "scripting language that compiles". That mean it has fast compilation and meant to use a "$lang run" shebang similar to Nim/Go/etc?
Interested to try this and compare with SPy: https://antocuni.eu/2025/10/29/inside-spy-part-1-motivations...
Thank you, yes please try out the language
Also I went through SPy's repo, and also looked at what they beat us in, we have advantages in certain areas. But so do they, so i've already planned new features to make our language even more powerful!
Did you look into Codon?
https://github.com/exaloop/codon