#social 2018-07-06

2018-07-06 UTC
fr33domlover2, fr33domlover3, fr33domlover4 and fr33domlover joined the channel
#
aaronpk is there documentation on Mastodon's use of activitystreams properties for CW posts?
#
nightpool[m] not specifically?
#
nightpool[m] we just use as:sensitive to mark the post as "sensitive" and then only show `as:summary` for sensitive posts
#
aaronpk ah there's a separate property that tells a consumer to only show the summary?
#
nightpool[m] yeah `as:sensitive`
#
nightpool[m] we discard summary otherwise bc, you know, microblogging :D
#
aaronpk what's the expected behavior of a consumer that doesn't recognize as:sensitive?
#
nightpool[m] hmm. not sure what you mean
#
nightpool[m] implementing as:sensitive is up to the consumer if they want it
#
aaronpk if someone starts a new AP implementation and doesn't know about this because it's not documented, they'll just display the whole post right?
#
nightpool[m] as:sensitive is documented
#
nightpool[m] it's an extension somewhere
#
nightpool[m] https://www.w3.org/wiki/Activity_Streams_extensions
#
nightpool[m] ```
#
nightpool[m] The sensitive property (a boolean) on an object indicates that some users may wish to apply discretion about viewing its content, whether due to nudity, violence, or any other likely aspects that viewers may be sensitive to. This is comparable to what is popularly called "NSFW" (Not Safe For Work) or "trigger warning" in some systems. Implementations may choose to hide content flagged with this property by default, exposed at user discretion.
#
nightpool[m] ```
#
aaronpk okay if someone doesn't find that...
#
aaronpk was there any thought to reasonable fallback behavior?
#
nightpool[m] we basically ported the ostatus semantics straight over to AP, including the fall-back behavior
#
nightpool[m] which people liked at the time we implemented CWs original for ostatus
#
aaronpk what's the fallback behavior then? showing the full contents ?
#
nightpool[m] s/original/originally
#
nightpool[m] Yeah, gnu social wanted to ignore cws all together
#
nightpool[m] and when pleroma implemented AP, they decided to show the summary but not hide the content, so it just becomes a title
#
aaronpk that's not fallback behavior, that's intentionally ignoring it
#
nightpool[m] ig
#
nightpool[m] the point stands that at the time, people we very much against a sensitiveContent kind of deal
#
nightpool[m] s/we/were
#
aaronpk I mean if a consumer doesn't know about the "as:sensitive" property, I guess they would end up showing the content?
#
nightpool[m] mm kind of depends
#
aaronpk i'm trying to figure out how best to translate this to mf2 markup that ends up with the same fallback behavior for consumers that don't understand the sensitive property
#
nightpool[m] if it's like an RSS-reader type of application, they might just show summary
#
nightpool[m] you know?
#
aaronpk example: my site doesn't have any concept of sensitive/CW, and check out the comment here: https://aaronparecki.com/2018/07/05/14/
#
nightpool[m] summary/content is the way we translate it for our mf2 markup as well as our open-graph tags
#
Loqi [Aaron Parecki] Which language would be the easiest for you to deploy and maintain going forward?
#
nightpool[m] isn't that the bug we were talking about?
#
nightpool[m] i have code for it I just forgot to push it
#
aaronpk if that "programming language musings" text was the "name" of the post, then my site would end up showing just that as if it were the name of the post, and it would hide the content altogether, which is one approach to fallback behavior
#
nightpool[m] that feels consistent with http://microformats.org/wiki/h-entry#p-name_of_a_note
#
Loqi [Tantek Çelik] h-entry is a simple, open format for episodic or datestamped content on the web. h-entry is often used with content intended to be syndicated, e.g. blog posts. h-entry is one of several open microformat standards suitable for embedding data in HTML. ...
#
aaronpk it sounds like that is not the fallback behavior that consumers get with activitypub though, so maybe that's not the right thing to do
#
nightpool[m] which can be "same as the title element" or "first sentence of the p-content/e-content"
#
nightpool[m] aaronpk: again, it depends on the intended UX
#
nightpool[m] most indieweb sites are much more of a full-featured blog ux
#
nightpool[m] in that case, it's natural to show the summary and link to the content
#
aaronpk right, my question is what is the intent when someone on mastodon uses that tag
#
nightpool[m] oh, that
#
aaronpk would they be surprised that the full comment text shows up on my blog?
#
nightpool[m] yes
#
nightpool[m] a mastodon user would expect any text under a content warning to require affirmative user consent before being able to see it
#
aaronpk so until we have some concept of this in h-entry, I think a reasonable solution is to set that summary text as the name of the post, since that would trigger (most) consumers to show just the name and requiring clickign through to see the content
#
nightpool[m] that sounds correct
#
aaronpk example screenshots: https://github.com/tootsuite/mastodon/issues/7926#issuecomment-402883007
#
Loqi [aaronpk] A change in the Microformats parsing spec from [a few months ago](https://github.com/microformats/microformats2-parsing/issues/6) reduced the cases where parsers would auto-generate a `name` property if there wasn't one in the original markup. Pr...
#
nightpool[m] summary <-> p-name sounds like the 100% correct mapping
#
nightpool[m] i guess i was maybe confused when you asked about fallback behaviors, because so far the types of people mastodon have been mainly federating with strongly disagree with the mastodon userbase on this point.
#
aaronpk oh interesting
#
aaronpk well by fallback I mean what does someone do when they don't even know about the extra property. intentionally ignoring it is a different story
#
aaronpk here was my attempt at changing it, tho I only found one of the files in this PR which is why I closed it. https://github.com/tootsuite/mastodon/pull/7925/files
#
nightpool[m] I have the changes locally
#
nightpool[m] after our discussion a couple days ago
#
nightpool[m] i just keep forgetting to push them
#
aaronpk looks like Gargron is about to push a fix
#
aaronpk https://github.com/tootsuite/mastodon/pull/7961/files
#
aaronpk I wanted to talk out this CW thing first tho, since currently that PR is just removing the p-name entirely
#
nightpool[m] i'm not familiar enough with the differences between p-name/p-summary to comment unforch
#
aaronpk well based on this conversation it sounds like the best/safest approach is to set p-name on the spoiler text
#
aaronpk and remove it from that .status__content containing element
#
aaronpk this is somewhat uncharted territory in h-entry since we don't have the concept of CW/spoiler text yet, which is why i'm talking about fallback behavior
#
nightpool[m] again, i try to defer to the UX expectations of the application at hand
#
aaronpk there's also a new proposal for post-type-discovery to better make use of the summary property https://github.com/tantek/post-type-discovery/issues/38
#
Loqi [manton] #38 add new step to use summary when there are unrecognized properties
#
aaronpk somewhat related
#
nightpool[m] and I think summary/content is semantic enough that the mental model users have won't be violated
#
aaronpk right now, in h-entry, summary is pretty much ignored in any case that the consumer understands the rest of the post's properties.
fr33domlover1 joined the channel
#
aaronpk anyway, if gargron merges this PR that removes the p-name entirely, we can start a new discussion about adding it back just to the CW preview text later.
fr33domlover2, fr33domlover3, fr33domlover4, fr33domlover, fr33domlover1, xmpp-social, vasilakisfil, vasilakisfil_, KjetilK, dlongley, tantek, JanKusanagi and jeremy joined the channel