Add byline metadata to constructor feeds

[reesericci]

This PR adds Byline metadata to constructor feeds to more richly describe authoring data.

[mdirolf]

This seems fine to me. Out of curiosity do you mind if I ask what the project you're working on is?

[reesericci]

I'm working on two things, a directory of crossword "columns" or publisher feeds, similar to the Podcast Index or the Apple Podcasts catalog, which requires feeds to actually publish crosswords with good metadata (like constructor authorship and .puz enclosures). There's essentially no standard for crossword syndication beyond vague notions of RSS, and I'd like to see that change. I'll probably write some adapters for AVCX, NYT and a few others, just to seed the index, but it's better if I can get it merged upstream!

This then allows my subsequent project, a macOS crossword app, to index, subscribe and display a variety of puzzles beyond the ones I personally implement or know about. There's no index of crossword puzzles, a standard to subscribe to and download puzzles from publishers, and no way to track constructors across publications. This ends up with apps like Black Ink implementing discovery by hand using a URL template which is quite brittle and bespoke:

Most puzzles are not distributed with a predictable URL pattern that Black Ink can automate the downloading of within the app

Yes, if a puzzle isn't publish on an expected date, it will not be playable without manually downloading the .puz

This is all in stark contrast to podcasts which have a well-defined syndication ecosystem that allows a wide variety of podcatchers to flourish!

I'm targeting Crosshare because it's open source here, and it has a large collection of open access puzzles. Hopefully, if Crosshare is indexed well and syndicates out, that will create enough of a network effect to encourage other publishers to also syndicate properly!