However, I feel like the key thing is that the tags need to be like 75% there before we let the community wild on them. The big piece of work right now is the fact that the tags are so wrong/missing etc. If you were to do this on a C# solution now, it’s probably only one tag change per solution, so it’s fast and easy (and probably more fun). So I feel like once we’ve got a good first pass in place for all languages, this should then get faster and faster. But that first pass being correct is really important. think.
I’m also very happy for Erik to spend time doing this for tracks without the maintainer support (although obviously that means he’s not doing other stuff, so keeping that to a minimum is good). Once we know which tracks maintainers do/don’t want to do, I’ll reach out to supermentors as the next step. Please just tag the issues where you don’t have the time to commit and then we’ll try and get others to help
There is an interesting case with construct:bitset tag, because it appeared in a code, where a charset function was used, which indeed returns a value of bitset! type, but I don’t know how the AI figured it out.
For me it was cross-checking the more obscure tags with the code. Some are easy to reference in the code like “construct:int”, but others like “implicit conversion” in some overly complicated spaghetti code take a lot of time.
For tracks without active maintainers (or no maintainers at all), is it possible for other members to post feedback in the respective github issues?
I appreciate that this may add overhead to Erik and the team, but it may at least help with the initial “pass” and minimize duplicates or wrong suggestions in future iterations?
I’m not sure about the process, as I can’t seem to directly edit comments I didn’t author on github – which appears to be what is being asked of us. Seems like it would be easier if these were just files in a repo.
I’m also unsure about the distinction between:
string interpolation and a template string or
a routine, a sub and subroutine
as these appear equivalent in raku.
With the raku seed exercises, most of the default labeling confusion seems to be about blended approaches and identifying something as using a class or method construct when it’s neither or multi-paradigm when it’s not.
You should be able to if you have write access to a repo. If you let @ErikSchierboom know where this isn’t working, he can look.
We’ll have a second phase coming soon where the community can start to add tags to tracks in a UI. In the interim, if you want to copy/paste and edit that’d be helpful too - please just make it clear what’s linked to what. Thanks!
