Hacker News Reader

Comment by culi | Hacker News Reader

Is it me or does it not seem very well thought out? Every example I've seen only has implementations in JavaScript and/or Python (without any mention of which version of this language the code is meant to run in. A critical piece of info for certain languages). I haven't seen any other languages nor a way to search by language. What a "string" means in one language can be completely different in another language. The primitive data types that the project assumes are not really supported across all programming languages.

Also if anyone hasn't already seen them, similar projects already exist and are more complete. E.g.

* https://rosettacode.org/

* https://programming-idioms.org/

* https://the-algorithms.com/

Not to mention LeetCode, CodeWars, Project Euler, Exercism can kinda serve the same role.

View on HN

> Is it me or does it not seem very well thought out?

Well, the first thing I expected to see when I clicked on a function was... the function.

It takes three or four clicks to get that, and the code is at the bottom of a fairly information dense page, where none of the information seems to have any relevance to using the function in your own code.

The other thing I expected to see is some discussion of the implementation, possibly allowing me to choose based on my needs (e.g. if I need to run a function 10,000 times each frame I might choose a different version than if I need to run it occasionally).

However, perhaps my expectations are wrong? Is this a tool for computer scientists to use in mapping the standard functions in various languages, rather than a reference resource for the working coder? The blog post doesn't go into any useful details, it's more like a standard marketing team announcement. They do mention it's part of wider initiative but it's not clear what that will be.

I do agree though that, especially in the age of AI autocomplete tools, I will never find this useful.

... actually, maybe it's a great resource for LLMs to learn from?

> Is this a tool for computer scientists to use in mapping the standard functions in various languages, rather than a reference resource for the working coder?

Neither of these. The video on https://www.wikifunctions.org/ gives a better explanation. If I understand it correctly, the project seems to be aimed at providing building blocks to more complex functions. Once the basic functions are defined in programming languages and described using natural language, you'll then be able to create new functions without writing any code, just by using existing blocks. Because each function is supposed to be described in multiple natural languages, you will be able to utilize them even if you don't speak English. Whatever you create you'll be able to embed on Wikipedia pages to perform calculations (e.g. automatic unit conversions) or string transformations.

https://mas.to/@vrandecic/110866878265998471

> I love Rosetta Code and we took a look at it. IIUC Rosetta Code does not allow to run the code, or to compose existing functions together to create new ones. Also, Rosetta Code's aims to have complete sets of implementations for their tasks, which is not our focus.

Wikimedia Foundation and letting their main product languish while they pursue badly-thought-out side projects? Name a more iconic duo.

Wikidata, Wiktionary, and Wiki Commons all heavily contribute to and compliment Wikipedia.

Wikinews is fine. It's not original but it's also not harmful to have more in that space.

Wikibooks and Wikiversity seem like they should be merged. Same for Wikisource and Wikiquote.

Wikivoyage is just not well made and there are other better projects out there

Wikispecies should just be a result of Wikidata, not a new project. Catalogue of Life (taxonomy) is much more thorough, rigorous, complete, and maintainable. No academic would use Wikispecies over COL. Misguided hobbyists might and that's a problem. And Open Tree of Life (phylogeny) already (thoroughly) serves the purpose of centralizing phylogenetic studies so there's really no unique thing Wikispecies provides and it gets outcompeted on the things it does provide.

Wikifunctions... I'm gonna be honest. I really think it's not well thought out and wish it didn't exist. I know a lot of people have already poured in a lot of effort into it and I always hate to shit on that, but as a long time Wikipedian I hate to see Wikimedia's name backing this up. Especially when I feel there are already other efforts out there that are more complete, better thought out, and better organized that would benefit from more contributors

Wikibooks and Wikiversity are pretty different. The former is about book-like or tutorial-like content (not just textbooks, but mostly any non-fictional content) that doesn't fit in the reference-work format of Wikipedia, whereas the latter is a catchall project for useful education-focused content. Wikisource hosts complete primary sources where allowed by copyright, while Wikiquote focuses on short well-known citations, including from modern works.

But yes I'm pretty sure that Wikispecies would not exist if it was proposed today, Wikidata has basically superseded it. The Wiktionary is a lot more popular but it could also end up being broadly superseded by Wikidata's lexical content.

I'd argue that sister projects such as Wikidata, Wikiversity, etc, are greatly useful and that the solution to improving them is not to neglect them, but to provide more resources to them.

Wikidata was launched by the same people launching Wikifunctions, in service of the same broader goals around abstraction of facts and semantic usage. A lot of similar design decisions are being made there, and it's arguably why Google.org invested in the Wikifunctions project (more and better programmatic interfaces to Wikidata = more semantic context across more languages that Google gets for free).

Im not super active in the wiki community anymore but last time I heard it being discussed, wikiuniversity seemed like something that caused more issues than it solved for wiki admins. Is that not the case anymore? Did it get better?

That's a byproduct of a lack of volunteer contributors rather than the project design itself.

Maybe so, but Wikipedia itself is fraying at the seams these days. Less and less Administrators every year, less and less committed editors, and more and more tech rot. See this comment https://news.ycombinator.com/item?id=38550441 for more context.

Reply on news.ycombinator.com