4 ms·
I find the article largely a troll. TL;DR: Implementations matte, and can profoundly handicap much significant and substantive discussion. This is long as it
by dredmorbius 5y ago
I find the article largely a troll.
TL;DR: Implementations matte, and can profoundly handicap much significant and substantive discussion.
This is long as it gets into the weeds of implementations and evolved practices, though illustrates some of the harms of CW policies and systems take to extremes or poorly considered and designed.
That said, there are implementations of content warnings which are very poorly designed and considered. One of these is on the microblogging platform Mastodon (and associated Fediverse platforms), which includes a "CW" option (identified in code as a "spoiler warning", giving some idea of original intent). Invoking a "CW" on a toot will result in the toot itself being collapsed, with the CW code highlighted above it, and a control to open/close the toot being shown ... if Javascript is enabled for the host in question.
This has come to be used both a general mechanism for concealing triggering material (legitimate) and as something of a substitute for a title or subject line. Worse, they're quite often used by people discussing topics near and dear their hearts which might be peculiar obsessions or interests but which have virtually no likelihood of being psychologically harmful to anyone --- technical or academic interests (software, language, environment, come to mind), or hobbies (woodworking, sewing, knitting, cooking/baking). Pulling a blanket over obscure but intense interests especially pains me. (Not everyone does this, though IMO far too many do.)
There are several disadvantages of these:
- Even for Mastodon users, CWs make some topics less accessible. In a medium which already elevates the trivial and deprecates substance (even without algorithmic bias), substantial discussion has even greater impediments. See: https://fosstodon.org/@alexbuzzbee/106666769704428696 https://fosstodon.org/@alexbuzzbee/106666769704428696
- There are now some tools for logged-in users which can expand a thread or even all (or selectively by keyword or regex filters) CWs. This helps, though the defaults mean that such tools will only be marginally useful.
- For anyone not logged into Mastodon, reading a thread by linking directly to it, which would generally indicate a strong interest in the content ... presents a view in which each individual toot is collapsed. Those must be inividually re-opened. My success rate in doing so without hitting any of the numerous nearby links and controls is about 80%, which means that on any discussion of 5 or more toots, there's a good chance I'll hit another control unintentionally, navigating off the page, and hence, have to start all over again. (This is much worse on Mobile where touch precision is lower than desktop.)
- There are other visibility controls, such as whether or not content is scoped as "public" (appears in the public timeline), "unlisted" (appears only to direct followers' feeds, though can be viewed by others), "followers only" (not visible except to direct followers), and "DM" (only visible to those directly mentioned). Those limits are somewhat countered by "boosts" which can expand the scope of public and unlisted toots.
- Mastodon has few discovery or search capabilities, though the one it does have (hashtags) require that toots be "Public" for them to be generally serachable, regardless of who is doing the searching. Ironically, hashtags can now be selectively filtered by most clients, though this requires the use of hashtags and a "Public" posting scope.
Again, the consequence is that a specific CW policy, say, one that requests all hard news or political discussion be CW'd, puts any such discussions at an extreme handicap in visibility. Note that such standards aren't uncommon, and may be applied across entire Mastodon instances. Scholar Social is a moderately-large (~3.5m members) instance, with one such policy, extending not just to "hard news" but "current events". See: https://scholar.social/about/more https://scholar.social/about/more
Several of the usability issues with CWs are illustrated where I discuss my own policies and where I generally won't use them (I don't have a complete aversion, and will CW particularly disturbing content). You'll have to open each individual toot to read the whole thing, however, after ensuring JS is enabled for the domain toot.cat:
https://toot.cat/@dredmorbius/105517193562512465 https://toot.cat/@dredmorbius/105517193562512465
I've repeatedly raised these issues. They're design flaws in Mastodon itself, though not interminable ones, and as such things go, not especially damaging. But it's a case where concerns for one group absolutely conflict with those of others, in a rather incommensurate way, and quite often for those whose interests are generally strongly alligned.
I'm not calling for no content controls at all --- Mastodon's mix of controls (user-level blocking, instance-level blocks of both users and instances, at several degrees) is actually generally quite positive, and provides a range of different "community standards" from which people can choose. (Some of those communities and standards I have absolutely no interest in associating with.) I belong to an instance which actively moderates others (toot.cat), based on criteria grounded in compassion, tolerance, and consideration for the most part (my own assessment), and is small enough that I can check with the administrator both on new suggestions and questioning existing ones. I've known the administrator for nearly a decade across other forums, there's mutual respect, understanding, and differences.
But the technical aspects of implementation are problematic in this case.