github.com
I decided to try to use it up over the weekend using (mostly) my phone and vibe coded a Prolog interpreter.
Now I'm seeing how far I can push it.
There have been over a dozen commits in the last 2 hours. If you are actually "vibe coding" a Prolog, then post the prompts you have used in the last 2 hours here.
If it helps to have a specific starting point, commit 7bbe652[0] is as good as any.
0 - https://github.com/nlothian/Vibe-Prolog/commit/7bbe652eaf0b0...
I don't remember when which part happened, but I've been gradually moving over to the tools in https://github.com/nlothian/Vibe-Prolog/tree/main/ai-tools for most of the development. Most of the coding itself there is done with kilocode.
Most of the early conversations don't seem to be available in Claude code now (is there a way to see complete history?), but here's a typical session:
> can this interpreter pass command line arguments to the prolog program?
[no]
> add these as missing to @FEATURES.md
[done]
> use `gh issue` to create an issue to address this.
> - Explain comprehensively what needs to be done > - Note that new test cases must be created > - Note that @FEATURES.md needs updating
> Do not provide estimates
That raised: https://github.com/nlothian/Vibe-Prolog/issues/75
Here's the PR for that: https://github.com/nlothian/Vibe-Prolog/pull/92/files
That PR was generated using my fix-issue-with-kilocode command. The source code for that is all in here: https://github.com/nlothian/Vibe-Prolog/tree/main/ai-tools.
The code reviews were done with Codex and Gemini, and responded to using my address_pr_comments command in that same directory.
This one's a straight port, paip-python's[0] Prolog interpreter. (Itself based on Norvig's Paradigms of Artificial Intelligence Programming book, originally in Lisp.)[1]
https://github.com/a-n-d-a-i/prolog.js
It was in the web ui, so I expected it to give me a code block, but it spun up a vm, set up a npm project, generated tests... ran them. I was quite surprised.
Your version is the opposite of mine, in a very good way. Your one is mostly tests! My version's tests are... well, you'll see ;)
Both are surprisingly short. 700-ish for mine, ~1200-ish for yours (as far as the actual interpreter goes), right?
That seems like a lot of bang per buck for something as powerful as a Prolog interpreter! I don't know very much about Prolog though, so maybe there's a lot of crucial parts missing here.
At any rate the original is a teaching device, and the book[1] goes into some length on the limitations of Prolog, both this version and in general.
[0] Original source in Python: https://github.com/dhconnelly/paip-python
[1] Original original source in Lisp: https://norvig.github.io/paip-lisp/#/chapter11
An easy way to find out if your implementation works is to try out the `likes.pl`[1] example found in the SWI-Prolog "Getting started quickly"[0] documentation.
0 - https://www.swi-prolog.org/pldoc/man?section=quickstart
1 - https://raw.githubusercontent.com/SWI-Prolog/swipl-devel/mas...
> It's not like it's a single shot generation or anything.
Since you've submitted a "Show HN: Vibe Prolog" and shared in the repo:
I was working on something else and as a side effect
accidently vibe coded a prolog interpretor on my phone over
the weekend.
Especially since there have been 82 commits in the span of 3 days.
EDIT:
And I am really interested in how you "vibe coded" the commit[0] which made this singular change:
- uv run pytest tests/ttest_additional_builtins.py -v
+ uv run pytest tests/test_additional_builtins.py -v
I don't think this is a violation of my (self-imposed) rules - it's supporting config, not code.
Before this[0] HN comment, I had seen no evidence of prompts used to affect change, only the commits pushed to GitHub. Maybe others had access to those prompts or I just missed them, either way I applaud you for having shown a sample of those used.
As to any determination of "violations", that is not my place. All I was hoping for was some evidence of "vibe coding" being employed. Without evidence, it is nigh on impossible for anyone to have confidence claims made are plausible outside of those making them.
> So there are two reasonable explanations:
> A) There is no way to show prompts used when "vibe coding".
> B) This claim of "vibe coding" a Prolog is disingenuous.
There's some more prompts here: https://github.com/nlothian/Vibe-Prolog/blob/main/docs/WORKF...
The author claims in the project README.md[0]:
I was working on something else and as a side effect
accidently vibe coded a prolog interpretor on my phone over
the weekend.
So there are two reasonable explanations:
A) There is no way to show prompts used when "vibe coding".
B) This claim of "vibe coding" a Prolog is disingenuous.
As for the rest of the code, it could be. Are there any obvious patterns you have found to be LLM generated?
In any case, my point is that a typo probably does not have to use LLM to fix that single typo.
So you can use LLMs but it is probably not used for everything, for example this typo.