Add "sorting by date" mode to "contributed music to groups" box

[towerofnix]

Not sure if this should display individual contiguous ranges (of years) or not.

For example, some of Erik Scheele's groups, showing just the range from earliest to latest:

• Official Discography (2010–2020)
• Fandom (2010–2021)
• Homestuck Gaiden (2010–2011)
• Beyond (2012–2016)
• Unofficial MSPA Fans (2012–2021)
• etc

And the same data, but individual contiguous ranges:

• Official Discography (2010–2013, 2016, 2019–2020)
• Fandom (2010–2012, 2016, 2018, 2021)
• Homestuck Gaiden (2010–2011)
• Beyond (2012, 2016)
• Unofficial MSPA Fans (2012, 2021)

It's important to consider this feature in the context of #308. Detailed chronology breakdowns per-group will be available by filtering, making the details here somewhat extraneous. However, this mode may also be used as a way to decide which groups you're interested in filtering by at the moment - in which case the extra details are useful, and directly reflected in the data you choose to filter.

Unlike "Sorting by duration", this does apply for artwork contributions as well as music. Client-side JS may need refactoring to support more than one of these sorts of controls being present on a single page (haven't investigated yet).

[towerofnix]

Another option is to display multiple items for each group - one for each continuous range of years:

• Official Discography (2010–2013) (sorted before Homestuck Gaiden because its range starts sooner)
• Homestuck Gaiden (2010–2011) (sorted before Fandom because its range ends sooner)
• Fandom (2010–2012)
• Beyond (2012)
• Official Discography (2016) (these three are sorted according to release dates within 2016)
• Fandom (2016) (above)
• Beyond (2016) (above)
• Fandom (2018)
• Official Discography (2019–2020)
• Fandom (2021)

But it's honestly kind of obtuse. Dividing by more discrete chunks of time is what the rest of the list is for, after all, and the details are clear once you've chosen what to filter by.

Worth keeping in mind, but probably not preferable above the two above.

[towerofnix]

Alright, thanks to Tangle in #hsmusic-chat, we've got a pretty fleshed-out design for this feautre now.

---

Concerns with the current group contributions info:

1. Sorting by duration or by count is not an especially useful default view.
   • Duration and count are both atemporal ranking kinds of placement: they provide information that simply isn't relevant if you don't care about comparing an artist's amoutn of work for one group to that of another. There is a casual use for that information, besides "for fun" — it can give a sense of what works an artwork might be especially known for, or prominent in the creation of — but this is mainly its only use. Details except for the very top can quickly turn into visual noise, decreasing the emotional value of the element as a whole (and encouraging you to just skip past all of it, all the time).
   • The element is not unreasonably huge, but it's still something that most people probably skip over when viewing an artist info page. If they aren't using, they have to skip over it, because it's always at the top of the tracks/artworks lists.
   • (Note: tracks and artworks do share the same default—both sort by count—so this isn't something a "sorting by date" would improve further on.)
2. Sorting by duration or by count isn't very good as a filtering control.
   • Not to say it's not good at all, but it tends to be poor for quickly accessing a select group you have in mind. Sure, Desynced is near the top of Rainy's list, but look for UMSPAF and SAHCon and the only cues you're getting for free are the group colors. Generally, since the lower end of the list is inherently more random, you have to do some searching. If you aren't aware of roughly how much the artist has contributed to the group in mind, sorting is absolutely random.
   • Sorting is moreover random per artist — each artist has contributed different amounts to different groups, so the list is always entirely personalized to them. At best you can usually find Fandom or Beyond closer to the top (since they include the totals of most quantities in individual other groups), but that's all you get. The most important groups for filtering (official / fandom / beyond) are at best semi-predictable across artists; the rest you can't build any sense across artists.
3. Presenting groups as completely discrete units, when there is almost always some degree of overlap, can be confusing to parse and understand.
   • This is a concern brought up by Makin in #issuefication-zone—thanks! Also what sparked Tangle's review and brainstorming.
   • The worst case is when an artist's contributed to a single fandom-oriented group, as is common for recent contributors to UMSPAF, for example: the list displays Contributed music to groups: Fandom (5 tracks), UMSPAF (5 tracks), giving the impression that the artist's contributed ten tracks total, when in fact they've only five!
   • But this can be broadly confusing anyway, ex. Rainy's top three are Fandom (74 tracks), Desynced (44 tracks), Deconreconstruction (21 tracks)—you can guess that 44 + 21 comes somewhere sort of around 74, but on closer inspection there are 11 tracks not accounted for. You have to parse the rest of the list to figure out what's what—UMSPAF (6 tracks), Stuck at Home Con (3 tracks) gets you all the way to 74—and there's no visual cue to assist. The numbers work out neatly for Rainy, reinforcing the apparent breakdown; but that's rare, as many artists have contributed tracks to albums that aren't (yet) categorized into specific groups besides Fandom.
   • The quantity-sorting modes are predominantly useful because they treat groups as discrete things, i.e. they don't regard overlap at all... but this is confusing at first glance, especially if you're parsing the list for any purpose besides getting detailed comparison data.

Tangle's proposal is to address all these concerns at once by using a spin on the "sorting by date" proposed here. Best to review the full discussion for details (#hsmusic-chat), but here's a summary of the takeaways:

• Make "sorting by date" the default view.
  • Chronological sorting reflects the structure of the actual lists! Quantity-ranking is something that "sorting by count" and "sorting by duration" uniquely do; otherwise, the entire page is sorted chronologically (and so are chronology links and most lists across the website). Quantity-ranking is sort of like a normal listing leaking into an artist info page's content, and default content at that. Although it may be useful (and is always neat trivia), it doesn't particularly fit.
  • Chronological sorting summarizes the content of the page. If you're interested in browsing the full list (or a filtered selection of it), "sorting by date" provides information that is relevant and useful: you get a sense of where/when certain contributions will show up.
  • Chronological sorting generally provides better stability across artists. Some groups have very long histories (UMSPAF spans about 11 years of history!), but most span a more limited timeframe. If you're looking for a given group, you likely know loosely when its music got started. "Sorting by date" provides you wtih simple numeric information that is consistent across artists (years are absolute, forward always means later), so you can quickly skip any entries that obviously predate the desired group.
  • Chronological sorting is great for filtering. Filtering necessarily means removing context; if the list of buttons you can click to filter is itself sorted chronologically, you can tell exactly what nearby context you're setting aside! Also, if you're choosing to filter by multiple groups, a chronological list will generally give you a sense of what things will overlap. If you're mostly curious what an artist was doing in 2020, entries including that year will generally clump together; even if they're not immediately nearby, the numeric data is still fairly easy to parse and find relevant ranges.
  • Making "sorting by date" the default leaves room for "sorting by count" and "sorting by duration" to focus harder on doing what they're good at, without bending to suit a more useful general-purpose view. And since the whole wiki places a great significance on chronology, "sorting by date" is inherently better suited to provide useful default information.
• Chunk groups under the main dividing groups of the wiki—Official Discography, Fandom, Beyond, Additional Tracks.
  • Broadly, these groups (defined in wiki-info.yaml by Divide Track Lists By Groups) represent how we conceptualize and the entire contents of the wiki at the very highest level. Apart from referencing-track lists, they also divide the "Random Pages" and "Artists - by Group" listings. And they're obviously constantly represented in data, marking the basic categorization of all albums and tracks.
  • Chunking by the dividing groups makes locating those dividing groups easier. These would always be sorted the same order on every artist info page (only skipping the completely empty chunks), so if an artist's contributed to Official Discography, that always places first, for example. This sort is arbitrary, but reinforced across the wiki, so it's a common cue. And since the dividing groups are the titles of chunks and everything else is indented, they're trivial to pick out of the crowd.
  • Chunking by the dividing groups makes locating sub-groups easier. They automatically put the main groups and all other groups on different (visual) levels, and there's no mixing up if SAHCon willl be indented or not; there are fewer albums to visually parse. Although some groups include material under two or more of the dividing groups (and these would be duplicately represented, though with different date data), most don't—you can always spot Desynced under Fandom, certainly never Official Discography. This makes for more consistency across artist pages, and generally easier and quicker parsing.
  • Chunking by the dividing groups provides data that is usually more useful. At a baseline, we categorize everything on the wiki under official/fandom/beyond; apart from "Albums - by Date", the dividing groups' info pages provide the basic "stories" which all albums on the music wiki fit into. So, chunking by these groups reflects those wiki-wide "stories" on artists' own pages—fitting the summary into the language the rest of the wiki uses, and making it more approachable/fun to explore.
  • Chunking isn't exactly ideal if you want a high-level summary of everything that an artist has contributed to—you have to mentally interweave the chunks—but that's a concession we're OK with making, mainly because the full contribution lists, when not filtered, already serve that purpose. (They're just more verbose about it.)
  • Chunking is incompatible with quantity-sorting. The point of sorting by quantity (count/duration) is to provide an absolute metric for comparing amounts. Chunking makes the final ranking muddier; if you want the data across groups, you have to keep mental track of the current maximum/minimum and cross-reference all chunks for the next largest/smallest value. It's awkward and inconvenient. Meanwhile, duplicate entries (groups which hold contributions under both Fandom and Beyond) are far less meaningful than in a chronological context, and outright problematic if an individual contribution counts as both fandom and beyond.

We'd like to introduce this feature at the same time as filtering (#308) — we (obviously) think they go well together. See https://github.com/hsmusic/hsmusic-wiki/issues/308#issuecomment-2067822462 for details.