#[tantek]ugh that sucks KevinMarks, and at first I thought you meant his Twitter handle whereas what you actually meant was his domain danhon.com was added to some Twitter linking blocklist
#[aciccarello]I've heard of similar malware flagging issues when sites hosted user content under the main domain.
geoffo, jacky, miklb, EncryptedGiraffe, barnaby, n8chz, bterry1, mro, jacky_, rvalue and [chrisaldrich] joined the channel; graycot[m] left the channel
#barnabypeople who have implemented any kind of moderation system for responses on your site: what states can each response be in?
#Loqi[marksuth] has 3 karma in this channel over the last year (14 in all channels)
#barnabyI’m considering having three states: pending, shown, and trusted. Incoming mentions from unknown sources are pending by default. When approving, they can either be shown with rel=nofollow added to all links to indicate that I don’t know the person, or shown as-is if I trust the source
#barnabythe nofollow rels would prevent random people who’ve replied to me but whom I don’t know from using that link as part of a webmention vouch for others, for example
#barnabybut I’m aware that I could also just be overcomplicating things
#barnabymaybe I’ll start with pending/approved on a post-by-post basis, and add in the nofollow stuff later based on an allow-list
#barnabygRegor: do you implement a block list or allow list in addition to post-by-post moderation?
#sknebelHow do you handle updates? I've been thinking a mode to "freeze" a response could be good, e.g. if the original link dies
#[snarfed]barnary: you'll probably also want rejected, so that you can keep them around (ie not literally delete them) even if you don't approve them
#barnaby[snarfed]: I was considering storing them with a rejected flag to prevent them from showing up again if someone re-sends the webmention, but thought it’d be cleaner to do that by adding the domain/author to a block list on rejection
#barnabyI suppose there’s a difference between spam responses I want to outright delete, and, say, abusive responses which I want to hide but keep as evidence
#barnabystruggling to think of other use-cases for storing rejected responses
#[snarfed]right, and even then, I'd recommend you still store them. audit trails, analytics, training your own spam filter, etc
#[snarfed]you'll also inevitably find one that you want to reject but not block its domain
#barnabysknebel: currently my site keeps archives of the HTML of each response, and on updates stores each measurably-different copy separately, so I always have the original for each version
#barnabynot sure I’m going to continue with that in the new version though
#barnaby[snarfed]: in the latter case, I’d add a moderation rule rejecting that particular URL
#barnabyfor my posts, I just set up a parallel “versions” table where on each significant change, I store a JSON blob of the previous version. I’m considering doing the same for responses, so I can see edit histories etc
#barnaby[snarfed]: I appreciate the feedback, I will definitely consider it!
#barnabystarted out with UX and social considerations but quickly devolved into technical details
#[aciccarello]barnaby, we also talked about different levels of trust in a recent Pacific HWC that were more granular than known/unknown. Let me see if there are some notes from that
#angeloaciccarello we were talking about marking your own posts with an audience. this is about marking an incoming post for display. i suppose you could combine the two and mark some incoming mentions as only viewable by certain audiences
#barnaby“don't have granularity as to quality of relation (link/know != approve)” sounds related to what I’ve been thinking about
#[aciccarello]Yeah, there's some room for different levels of trust between "I 100% trust" and "Block this person"
#[KevinMarks]I think you mostly want to keep them, but may want to expurgate certain kinds of posts with embedded criminality
#barnaby[KevinMarks]: agreed, I’m definitely going to have a “block and delete” option for extreme cases
#capjamesgHow do we make setting up an IndieWeb Website as easy as creating a TikTok account?
#barnabyre different levels of trust: yes, and I want to avoid the case where (assuming wide adoption of /vouch) someone leaving a legitimate comment on my site allows them to spam my friend’s site
#[snarfed]capjamesg micro.blog plus integrated domain registration that handles DNS etc automatically
#[snarfed]sorry, didn't mean to imply that it didn't
#barnabyeven then it’s not as easy as it requires paying for the domain, but other than just using subdomains (which micro.blog, known, tumblr etc do already) that’s the closest we’re likely to get any time soon
#[snarfed]true. arguably it's a feature, paying users are a better business model than ads, etc. but looking strictly at onboarding, yes, it's friction
#[snarfed]I vaguely remember there was some TLD (European?) that gave out free pay-level domains. couldn't find it last time I looked though
#barnabyagreed, IMO it’s only really an issue when the question is framed as “how do we make it as easy as tiktok’
#barnabyI wonder how viable it’d be to campaign, say, to the EU for having a free personal domain being a right for every EU citizen
#barnabysounds like a pipe dream, but as we’re seeing more and more legislation focused on breaking up social silos and (in theory) enforcing data portability and interoperability, it might be possible to make a case for it
#LoqiModeration is the process of holding comments for review by a human, and sometimes a source of frustration when a comment is written on another site that the commenter has no control over https://indieweb.org/moderation
#barnabythat was the first thing I looked at, didn’t find much of interest and wanted to hear from people with experience implementing and using moderation on their sites
#barnabyso I’d encourage anyone who does implement some sort of moderation (e.g. gRegor) to add themselves to an Indieweb Examples section on that page ;)
#[tantek]IMO it's worth having a higher level conversation of good defaults for showing reactions / replies in the first place, with moderation as one possible approach
#[tantek]IMO we have to figure out UX ways to make receiving webmentions and handling them as close to fully "automatic" as possible, and *not* require moderation in "normal" cases
#barnabyagreed, I’m currently in the process of reqorking my site and want to start off with a basic moderation system so it’s quick and easy for me to approve/reject responses, which isn’t the case currently
#barnabyso I was asking from the POV of what’s the simplest, quickest thing I can throw together which other people have found useful, while keeping in mind more advanced solutions to implement later e.g. allow lists, vouch etc
#Loqifriendly reminder barnaby, that's a lot of dev jargon! implement... can you move to #indieweb-dev?
#barnabyI’m satisfied with the indieweb-specific answers I got for the moment, but I definitely want to add to /moderation at some point with additional examples
#gRegorMine isn't really moderation. I have a UI I can manually set a response as private, but the majority of the time visibility is set automatically if there's an error processing the response.
#gRegorOtherwise if processing the response works, it's public by default.
#barnabyhave you ever used the UI to hide a response you didn’t want shown?
#gRegorMaybe. I can't remember. If so, less than a handful of times and it was probably a spammy Twitter reply, so I just deleted the response from my site.
#gRegorI can check to see how many I have set as private now
#LoqiIt looks like we don't have a page for "ProcessWire Webmention" yet. Would you like to create it? (Or just say "ProcessWire Webmention is ____", a sentence describing the term)
#[tantek]Consider how you don't really bother “moderating” comments on GH issues you file in general
#gRegorMost of my moderation would be on that wiki page or the module's documentation ^.
#barnabywith GH issues you benefit from their anti-spam infrastructure
#gRegorCan expand if there's certain things you're looking for
#[snarfed]agreed, fwiw "never have to think about it" is where I've been for a long time. I run Akismet on normal comments and auto-approve wms. very very rarely have to manually moderate anything
#[snarfed]I'm not usually in conversations or topics that attract trolls though, fwiw
#[KevinMarks]what should I linkshare for leo's crew?
#[KevinMarks]I could send them to sane, but it may be premature