handling content warnings, tagging, and other things of this nature
Alyaza Birze
Public
Seen by 28
i think most of us have at least one issue with how the Fediverse technically or culturally tends to handle content warnings, tagging, etc.; unfortunately, as of right now we'll inheriting most or all of those issues while we have to rely on frontends that we aren't developing ourselves. so, let's talk over those issues.
i am, for the most part, a content warning/tagging/read more/etc. purist. if a thing says it is for content warnings, my expectation is that it should be for content warnings alone―only in a handful of weird circumstances should these any of these really meet. this was how Cohost did things: content warnings were separate from tagging, and tagging and content warnings were separate from read mores, and so on. this allowed for lots of granularity as the situation demanded: you could hide things with any combination of tagging, content warning, and read mores and it was fairly intuitive all around as a system (even if there were some downstream issues like cultural tag collisions).
unfortunately, most Fediverse software and frontends blend all of these to one degree or another; for most intents and purposes the "content warning" field is closer to a `<details>` tag which handles (or is expected to handle) all of these things and more. i consider this exceedingly bad for users and exceedingly bad for the software itself, and i think it's going to cause us nothing but headaches if we don't have an ability to disambiguate them.
in short, as i put it in short earlier on Discord:
content warnings, readmores, summaries, and failsafes for "long post display weird" or "thing that does not display well" etc. are all discrete use-cases to me and simply put: probably should not all rely on the same field as load-bearing infrastructure
if it is at all possible for us to change that, my feeling is that we should do it ASAP.
sirocyl Fri 4 Oct 2024 3:17AM
@Maya As in, require using "#CW: US politics #CW: Trump". Tag synonyms (the "#CW: uspol #CW: US politics" thing) should be permitted.
Shel Fri 4 Oct 2024 1:11AM
I have written my opinions on best practices as a twenty page essay.
Maya Fri 4 Oct 2024 2:30AM
@Shel This is such a valuable writeup of the problem and how the different implementations came to be.
nora Fri 4 Oct 2024 2:56AM
Commenting, here, only on the recommendations made in this post:
As to suggestion 2, about tags, I agree. We already have this - yay!
As to suggestion 3, as far as it relates to text filtering, I agree. We have this feature and could easily add a policy amounting to "make sure to use these keywords in the post somewhere or at the bottom".
Suggestion 4 doesn't apply to us (but I agree with it.)
As to suggestion 5, as far as it relates to CWs specifically, I agree. As it relates to tag etiquette, I think we could easily implement such a policy, but I don't think that "needed for your own accessibility and filtering needs" and "advocat[ing] for an imagined other" are the only two cases; any such policy should address the case of a person who does not need a particular tag advocating for some real friend or relation who does.
As to suggestion 6, I agree. If policy can be made from that, I would likely be in favor, but I don't see how to do it.
As to suggestion 7, I agree wholeheartedly.
As to suggestion 1 and the part of suggestion 3 related to when CWs should not be used, I disagree, particularly based on the principle that it is inordinately intrusive to tell our users that they are not permitted to use certain CWs, except when that use is, in and of itself, an act of bigotry or otherwise breaks our rules. I remain unconvinced that people putting CWs on their posts that we ourselves wouldn't add is actually antisocial or negative for the network's culture.
Maya Fri 4 Oct 2024 3:55PM
To expand, I'm nervous about the metric of CWing things that only most people find shocking. People naturally use their in-groups for that sort of metric, and I can think about all sorts of situations where that can be a source of cultural friction that causes moderation problems or chases people away.
isomorphism Fri 4 Oct 2024 10:49PM
Shel's written a lot and more eloquently than I could. I'm mostly in line with what she wrote and with her recommendations. I've got PTSD from some recent stuff, and I'm not expecting a place that bends everything to accommodate me, or to never engage with things that could be tough to see, just a system that allows me some capacity to curate and calibrate for myself. Cohost worked well for me on that front.
Maya · Fri 4 Oct 2024 2:22AM
@sirocyl Disabling search for CW tags is key if we want people to use them.
By no combining, do you mean "#CW: us politics, Trump" would need to be "#CW: us politics, #CW: Trump"? Or do you mean that "#CW: us politics" and "#CW: uspol" wouldn't be treated as the same tag for filtering? (This latter assumes that we do end up implementing some sort of league wide tag wrangling system, since no instance software currently support that AFAIK.)
I like the idea of using a "namespace" with the tag system as a quick means of implementation, though eventually we probably would want the front-end software to make it hard to accidentally enter a CW as a normal tag.