Cadiac / sanuli

A finnish version of a popular word guessing game, written in Rust.
https://sanuli.fi
MIT License
64 stars 6 forks source link

Fabricated words in the daily puzzle #29

Closed nikcorg closed 1 month ago

nikcorg commented 1 month ago

Apologies for using this channel for feedback, but I couldn't find any other ones.

Recently the solution to the daily puzzle has on multiple occasions been a word not included in the Kielitoimiston Sanakirja wordlist. Today's word (12.7.24, to avoid spoilers) hasn't even a (Finnish language) definition discoverable by online search.

[Edit. In the end, I did find today's word in Kielitoimiston Murresanakirja. Maybe the change I'm wishing for is if the game would link to the right dictionary, instead of always to Kielitoimiston Sanakirja.]

I can understand the desire or need to include words beyond the original corpus, but, if I may, I would like to propose that all included words should have a linkable-to definition, preferably by a dictionary.

Thanks for reading and, despite this minor gripe, many thanks for the game.

Cadiac commented 1 month ago

Words on the daily puzzle have been picked mostly randomly from the full word list, which started off from the Kotus "Nykysuomen sanalista" in 2022, but has since been adjusted based on the feedback received, as people mostly wanted more pretty reasonable spoken finnish words to be added that were missing from the original dictionary material.

As of today, I've received nearly 27k requests to alter the word list through the Google Forms, but nowadays I have little interest in actually maintaining it as that work was quite labor intensive and this game is/was a small hobby project of a single person. Nowadays I just skim through the feedback every now and then, sometimes changing something but this project is currently on pretty inactive maintenance mode. I plan to keep this game online for the time being, as there's still a reasonable number of active players and hosting this only costs me the domain.

The daily word of 12.7.2024 (kuiru) was indeed a bad one, and while I've attempted to manually review the list https://github.com/Cadiac/sanuli/blob/master/src/bin/generate-daily-words.rs script generates (while removing duplicates that have already been the sanuli of the day) this must've slipped through last time I extended the list of daily words in January. Based on the Google Forms feedback this upset many users, and I should be a bit more careful in generating the list for päivän sanuli.

I'll probably drop "kuiru" from the word list as it has a pretty obscure meaning, it was added on the relatively early days 2022-01-23 after someone requested it, explaining it being a word for a type of a boat.

The link to a dictionary is a "best effort" lazy solution - just always linking to kielitoimiston sanakirja with that word as the query parameter. Support for linking to alternative dictionaries wouldn't be difficult to implement, but gathering the dictionary link to use for all ~8k words on the full-words.txt would take a lot of half manual work that I'm at least unwilling to do at this time.

Thanks for the feedback anyways!

nikcorg commented 1 month ago

Thanks for responding.