ArtisanChat Documentation
Rules & Guidelines
- Terms of Service
- Server Rules
- Content Warnings
Funding
- Patreon
- Ko-Fi
- Financial Reporting
- Long-Term Goals
Moderation
- The Team
- How To Moderate
The Technical Side
- Technical Specs
- Disaster Recovery Plan
Everything about ArtisanChat (eventually)
Content warnings are a complicated subject, but after you see them in use for a little while you'll realize how much more pleasant they can make the experience. And, counter-intuitively, how much wider your engagement will be if you use them properly. Tell me a terrible thing without warning, or show me a gory news picture in the middle of my feed and I'll mute you and never see you again.
But tell me an important story with an appropriate CW or a marked-sensitive gory image and let me approach it on my own terms, and I will listen to what you have to say now and in the future... and I might even boost your post for better reach.
Not all instances adhere to the same CW policy, so if you're in the Federated timeline you may see posts that break our rules. If it's an egregious problem, report it – if it's not against the remote site's rules we may choose to silence that site to keep that content off the Federated.
If it's content that's not against the rules but that you never ever want to see, use ⚙️ > Filters to remove it from your timeline. You can temporarily filter things, like a sporting or conference event that is too “noisy” but that you don't care about seeing a little traffic from after the fact. Or you can drop things forever, like a certain daily word game.
You can mute or block a specific user, or even their entire instance. Go to their profile and click the stacked-dots menu. Blocking the entire instance is the very last menu item. (But if it's an instance that should be blocked at our instance level, tell @Artifex!)
On the flip side, if you find yourself opening every content warning you see, you may want to set them to be open by default. ⚙️ > Preferences > Appearance > Sensitive Content will give you the choice to always expand the posts. Choosing to always display sensitive images is another, separate choice.
You can also choose to never display sensitive images (they'll always be click-to-open) if you prefer.
Give serious thought to your followers when boosting a post without a CW – is it what they've come to expect from you? It's perfectly valid to ask someone “can you please post a version of this with a CW so I can boost it?”
The first reaction most people have to being asked to use a CW is “I am not going to self-censor!” That's not the right outlook, at least for a good-faith request. The people who don't want to see your content at all will filter on keywords or mute you. But if you want to reach people who find your subject heavy but still want to engage with it and you, use a CW. People will read you more if you respect their boundaries.
Proper CW use is surprisingly non-intrusive – readers quickly develop the habit of clicking or skipping without giving things much thought.
It's the tragedy of the commons: yes, your topic is so very important that it can't hide behind a CW. So is mine, and the next person's, and so on. We all have to make compromises.
Media posts don't need a CW if the text isn't sensitive – that just adds a second click for readers to get to it. Make sure the text gives enough information for the reader to know whether to click or not.
If you mention a frequently-CW'd topic in your post but it's no more of a passing mention that a CW itself would be, a CW is not necessary.
If someone is reading posts on your profile, or as part of a thread, things don't hit the same as when you're interspersed in the public feed with the cat pictures and whatnot. You should still CW your lewds and whatnot, but it's more “saying it in your living room with the front window open” than “yelling in the public park.” It'll cut down on your readers' willingness to boost, though, so it's a tradeoff.
If someone wants to yell at you about a controversial post that was clearly marked in the CW, telling them, “You didn't have to open it,” is a pretty ironclad defense. At least, provided your CW was neutral and not itself baiting, and provided your post isn't breaking any other rules.
Clients vary widely in their handling of long posts. Even a “standard” 500-character post can be very tall (yes, you could make a post that's 400-some carriage returns. DO NOT). Putting a CW as a subject line gives people a way to collapse your post even if no other method is built into their client.
There are some fairly standard CW tags; even if you're just marking an image sensitive, try to have the relevant word in the text somewhere so filters can catch it.
politics (uspol, depol, eupol etc. are also common) – The original CW. Before you say “but everything is political,” understand that “CW your politics” came from the timing of Mastodon's birth: in fall 2016. We had all come from Twitter utterly sick of an ugly US presidential campaign followed by an even uglier result. Mastodon was the only social-media place to get away from a constant drumbeat of doom, and randos showing up in your mentions to smirk and gloat. Some people still wanted to talk politics, but we also wanted not to have our goofy posts interspersed with gut punches. CW the heavy stuff, we beg you.
covid – We've all been through some pretty deep trauma, personally and as a society, and while like politics the pandemic is all-pervasive, people still like a little buffer before anything substantial.
death – Again, we've had too much of it lately. Fictional death, pet death, even implied animal death is often seen. If your post could ruin someone's whole day, why wouldn't you CW it?
racism , bigotry, sexism, homophobia, transphobia, ad infinitum – As with politics, this is complicated. I CW talking about racism because maybe Black people don't need to hear yet another white person literally named Karen's opinion on it. A POC may not because that's just another day. On the other hand I, as a woman, choose to CW sexism/misogyny because, as I say above, I want to buffer it and because I believe it leads to more engagement, not less. YMMV. (PS: a CW for racism is not cover to actually do a racism.)
body horror, tattoos, injury, etc. – If your post is likely to inspire “cannot unsee!” you probably need to mark that image sensitive. Respect people who struggle with self-harm. Understand that some people are squeamish about tattoos; fresh ones or closeups almost always deserve mark-sensitive.
alcohol or just alc – Alcoholics and recovering alcoholics appreciate the buffer of a CW (or the filter-ability).
drug, drugs, medication, meds – Same as the above, with the additional note that you should help out the folks who use it to filter by using the first word for recreational types and the second for medicinal ones, and watch your terminology in the text as well so you don't wind up a false positive.
food – This may seem surprising, but there's a strong “CW your food!” tradition. It's not mandatory, but it's helpful enough to those people who struggle with eating disorders or inadequate nutrition that most of us do it. Many people add veg or meat qualifiers as well.
weight – Relatedly, talking about weight is complicated for people with eating disorders, health issues associated with rapid weight gain/loss, body-image issues – the list goes on.
eye contact or ec – Another surprise to some people, but folks on the autism spectrum and the like often find eye contact very off-putting. Some people even CW animal or cartoon eye contact.
mh or mental health – This CW is often less for the reader's protection than the writer's: “I'm about to be real here, so please be kind.” Many people use mh (+) or or mh (-) to prepare the reader for the tone of their post (this is true of many other CW topics).
There are a lot of less-obvious things that CWs can be used for.
spoiler or spoilers – Make sure you put what the spoiler is for, obviously.
emoji – Hover over an emoji to see what a screenreader will say, often preceded with “graphic.” Now imagine what it will sound like when you have a string of rollingonthefloorlaughing emoji. Saying something like “lots of emoji” or “heavy emoji use” will let folks with screenreaders decide if they have time for that kind of thing. This is generally not required for the special case of...
Wordle – This is both a spoiler and emoji-heavy, so most people CW theirs.
Jokes – when the feature was first introduced, the timeline was just a wall of knock-knock jokes. It was so stupid, and hilarious. You will still see it fairly often, including “made you look” type posts that are things like “some” followed by Smash Mouth inside the CW.
Long post subjects, as discussed above
The current hotness – As I'm writing this, the Twitter migration is in full swing, and all anyone new and old wants to talk about is Mastodon itself. This is very much like early-Mastodon days, so a lot of people have taken to tagging meta discussion as meta to cut down on its domination of the timeline so the real community can get built. Sure, people can filter, but maybe they want to get some of the info.
In a past life, I ran IT for a number of different places, most recently including an airline and a bank. It's been awhile, but I remember enough to know disaster recovery plans are important.
What happens if our data centers burn down? ArtiChat is currently hosted on MastoHost, which in turn contracts for space in data centers and whatnot. One hopes they have their own redundancy and recovery plans, of course, but to a certain extent we have to take those on trust.
MastoHost offers external backups, which I take advantage of, because I don't trust anybody that much. Worst case, I point our DNS record to wherever we set up on a new, un-burnt hosting service. We'd lose up to a week of posts and new accounts and whatnot, but as catastrophic failures go it could be worse.
What if we get compromised? That gets a little more complicated but again, absolute worst case we start over from an uncompromised backup. (And I get to spend a lot of time sending out notifications to all y'all.) I keep three generations of offsite backups, all in one place and nowhere else. And I use 2FA everywhere, and am one of the most deeply suspicious people you'll meet... won't say I can't be phished, but at least they'll have to work at it.
What if I get hit by a beer truck? I am recording everything about ArtiChat with the intention of handing it off to some kind of board/organization anyway. All the passwords are in my archive with the password manager's deadman-switch thingy so if I drop dead my husband can manage the handover. One of MastodonART's moderators has admin access to keep the thing running in the meantime; mART will take over if I haven't put together a board here by then.
(Archived from this thread)
There's been a lot of drama around how sites choose to block other sites, and some concern (mostly insincere) about mods having to look at problematic sites all day, so here's a little thread on how I do it.
The vast majority of sites in our blocklist just say “edgelord,” and they're detected by a bot, and I have never looked at them. Before you decide that's awful and probably unfair to those poor sites, let's walk through that.
The Federater reports are a prettier front end on a tool we've had access to almost since day one. That tool surfaces new sites for blocking almost continuously, and most of them are tiny. We'd probably never have contact with them, but it's useful to pre-emptively block in case of a dogpiling incident: fewer holes for moles to pop out of makes whac-a-mole easier.
Grabbing the most recent report, it's a site the underlying tool has reported on before so it's already blocked here.
https://federater.github.io/crucible-world.html
It's an easy block: every single site on there is throwing up danger signs. Even without the emoji, I would instablock upon seeing that poast is the most popular site with the admin(s). That's one of the edgelord hubs. Zero normal sites in there at all.
https://federater.github.io/piazza-today.html
This one looks fine as admin, awful on federated. Now, you can't always judge a site on its fedline: some admins don't ever look at it (I peek in on ours now and then), or have signed up with relays without a full understanding of what they're getting – in those cases, just because horrific sites show up, it doesn't mean anybody on that site is actually following those accounts, or is even aware it's showing on federated.
Federating with the bad guys is still bad (there are some “secondhand smoke” effects), but fixable. When a site blocks based on “federating with the usual suspects,” it doesn't usually mean they just show on the fedline, but that there's interaction, a reason to think they're fellow-travellers.
In this case, it's a single-user site so that's all him. (And yep, I checked the site before blocking, because it still might be a false positive.)
We're not big enough to show up in public Federater reporting yet, but I know what we look like on the back end right now: popular-with-admin top sites are mastodon.art and wirebird.com (our status page). Next time we get sampled it will probably be a little more spicy, since I've been interacting with other admins recently, including a couple I later suspended (for not handling the spamming I was asking about).
Our top sites in federated are mastodon.online, mas.to, mastodon.lol, wandering.shop... and mastodonapp.uk. That last one is problematic, in a non-obvious way. A lot of other sites have defederated from them, and it's probably one of the first orders of business our board will take up, when we have one. A lot of suspensions are easy-peasy, but a lot also take manual consideration.
And that's about all there is to it.
(correction: since I started writing that this afternoon, we showed up because I posted to the fediblock hashtag. Still clean! https://federater.github.io/artisan-chat.html )