quote + boost

[triplefox]

This turned out to be a major Twitter feature: being able to inline the text of a retweet while adding your own. It was done by manually copy-pasting the text before a feature was developed. How to go about implementing this, or an equivalent? How would it work with our visibility rules?

[clayote]

1. When a toot contains a link to another toot, embed a preview of the latter in the former
2. Some kind of interface affordance, button or menu entry idk, to automagically copypaste a link to the toot your mouse is on into the toot editor, then put your cursor before that
3. Notification when someone quotes your toot this way

[Gargron]

Could be something like OEmbed, then the same mechanism could be used for embedding youtube/soundcloud/GNU social statuses/anything that supports OEmbed.

[hikari-no-yume]

How would this work with federation? Has anything in the wider network implemented this?

[hikari-no-yume]

quitter.no apparently can even quote tweets (that is, from Twitter). Dunno if that's standard.

[Gargron]

Okay, so GNU social quoting tweets is the same mechanism as any link (it does an OEmbed), but for actual fediverse statuses it seems to use a slightly different mechanism, and I'm thinking I might want to specialize it too, since it'd be useful to create notifications about quotes, and also have a little more control about displaying them in the UI.

[ghost]

i would suggest making a quote tweet be noticeable as a reply to the post, that also has the link and OEmbed view of the toot. this would happen even if the user just copied a link to the toot, ideally, to make it so people can't effectively do what not implementing quote-toots would do, namely, allow people to know when people are talking about them.

[pief]

Would love to see this, too! +1

[Swipe650]

Any progress on this? Would love to see this implemented.

[ghost]

@Swipe650 yes me too, it would be a great new user-oriented feature. No milestone though.

[El-Gavy]

+1 !

[BasixKOR]

It is what I need.

[clayote]

I would also be satisfied with some way to make it "stick" when I boost something and then post about it -- that is, make sure the boosted post and mine will appear next to one another in the timeline.

[fenarinarsa]

This is actually a major feature on Twitter and it's missing badly on Mastodon.
I hope it will be implemented soon :)

[Cassolotl]

I like quote-boosting and no harassment, so I'm wondering whether experienced admins would be willing to share their thoughts on the possibility of quote-boosting being allowed, and there being instance rules that say "no using quote-boosts for evil" or something like that?

Edit: This is because I've heard several times now some rumours that the reason quote-boosts haven't been implemented is because it's likely to be used for harassment and other bad stuff - similar to the motivation of only being able to use the search for hashtags, usernames and toot URLs.

[thiagomgd]

I love and hate the quote option on the birdsite. Why:

Love: I can share AND comment at the same time. Of course, boosting and then replying would have kind of the same result, but anyway...

Hate: people use that to spread the flame wars. Instead of replying the poster, they share / comment, just to make their followers be able to harass the original poster.

So I have a few considerations:
Instances should be able to block it
Users should be able to block it (let's say I don't want anyone quoting me)
And of course, each instance would add some rules about that.

And at last, of course, we should see the embed/quoted toot, as we do on that other site.

[fsnk]

I too would like this feature. I feel like it adds value to the conversation to know why someone in my timeline is boosting something. If they agree or disagree or what their thoughts are or it's their friend's artwork or whatever. I also think that embedding the boosted toot makes for a more streamlined experience.

Something I dislike about how this is implemented in other platforms is that the comment becomes the primary status for replies, which can create confusion in threading. Especially if people are quote boosting replies to quote boosted replies to quote boosted replies.

Regarding the harassment issue, I've definitely seen this used on other sites to start dogpiles, but I'm not sure that the bad use of the feature outweighs the value it adds in giving context and enabling conversations. I've seen Mastodon users screenshot other users' posts to criticize and denigrate already; all it needs is a link to the original status and that behavior is essentially duplicated. And the original user never knows it's happening, unlike a quote boost where they should get a notification.

Having an automated feature would make harassment easier by cutting out the screenshot/image add/link paste actions, but not having the feature doesn't stop it. (And that technique is still used on other social media sites that do have the ability to quote boost when harassers want to obscure where the dogpile is coming from.)

So I guess the thing is, how to implement this feature in a way that gives the original poster some control over the boosting? Maybe give users the ability to remove their status from the quote boost, leaving a comment that was intended to provoke harassment without context? (Much like deleting a quoted tweet does on Twitter.)

I don't know. I would definitely like some kind of feature that allows me to attach a comment to a boost though.

[nolanlawson]

I agree with @thiagomgd that this feature is a mixed bag. In fact, overall I'd rather not have it, or have it turned off by default.

Reason being: this feature is often misused for dog-piling and abuse. When it was first introduced on Twitter, I saw it quickly appropriated for the purpose of directing one's followers against another user. Frequently this spirals out of control, with endless quote-tweeting (such that you have to click multiple times to even follow the kaleidoscoping conversation), and tweets are often taken out of context (especially if it makes for a better punchline at the other person's expense).

Quote-tweeting can be used for good, but in my experience it's more often used for nastiness. I call this the "Hey everybody, get a load of THIS jerk!" phenomenon.

Of course you can already do this with a screenshot, but at least with screenshots you're not notifying the other person that you're mocking/disparaging them (which can escalate the situation). Plus quote-tweeting makes this kind of harassment easier to do, and affordances matter.

I don't always agree with Paul Graham, but his assessment of this feature rings true with my experiences.

[Cassolotl]

I don't feel as strongly about quote posts as I did in 2018. Personally, I am not a fan, but there is clearly a lot of demand for it. We're considering it.

If we did do it we'd like to make it something you can opt out of, in a similar way to how we plan to allow disabling replies. It's not entirely trivial.

--Gargron

Perhaps this issue should be reopened?

[LeftIndy]

I don't feel as strongly about quote posts as I did in 2018. Personally, I am not a fan, but there is clearly a lot of demand for it. We're considering it.
If we did do it we'd like to make it something you can opt out of, in a similar way to how we plan to allow disabling replies. It's not entirely trivial.

--Gargron

Perhaps this issue should be reopened?

I concur. The issue can and should be re-opened.

[forenta]

Patch for Quoted Posts on Mastodon is available on Gitea.

[Cassolotl]

@forenta Oooo this looks niiiiiice!

[bleonard252]

Worth considering for any implementation of this feature (if it hasn't already been mentioned): make sure that the OP is notified (if it isn't yourself, of course) to help prevent talking behind peoples' backs. For notifications, this should be counted as a different event so the OP can turn off quote-boost notifications if they so choose.

[Cassolotl]

Someone:

I miss quote tweeting the most. I rarely abused it and I think the argument it was primarily used for abuse both devalues the real value of why USERS invented the quote tweet to begin with and gives too much power to the so-called abuse that could be better solved with better abuse/block tools.

Someone QTs them to say:

We’re just reinventing it from the grassroots once again. Clients like Ice Cubes and Ivory show previews of Mastodon posts when you include the URL in your quote post. It’s already happening and Mastodon’s official web interface will have to catch up. They lost their chance to define the quote-post on Mastodon by refusing to do it all for highly dubious “safety” reasons.

I am concerned that Mastodon developers aren't keeping up with what people need and are implementing for themselves.

I think (and I might be wrong) that if Mastodon implemented QTs with safety features that people would want/need for QTs to be comfortable and good overall, such as individual per-post QT opt-in (or per-user QT opt-in), clients would follow vanilla Mastodon's lead.

@Gargron, is it time to reopen this issue yet? It's starting to feel more urgent.

[trwnh]

Clients like Ice Cubes and Ivory show previews of Mastodon posts when you include the URL in your quote post

vanilla mastodon does this too, this isn't the point of contention with "quotes". the preview is not the problem. if people end up copypasting links then that's not really the same issue -- linking to web pages and other web resources is a perfectly normal and uncontroversial action that you've been able to do since forever ago. if these links contain preview information, then you can show that preview information.

the problem begins when you start to expect additional behaviors. should posting a link always notify the source of that link? should replying to a post with a link auto-tag the author of the link? what is the social or communicative purpose of the link? questions like these should be considered very carefully by anyone wanting to build out more affordances on a technical level.

put another way: what defines a "quote" to you? i'd break it down along the following lines:

• that it is embedded is a given, or else it would not be a "quote", it would just be a "link"
• notifying the author should be an active decision. currently to do this you would have to explicitly mention the person you are referring to. if support for Webmentions were added, you should explicitly decide to send Webmentions, and the other party can explicitly decide whether they want to receive them.
• should it be part of the same thread? should it be considered a reply?

that last point in particular is something that i've thought about recently and come to the conclusion that a "quote" is essentially one of two things: either it is a "loud reply" or it is a "breakout thread".

a "loud reply" is, in effect, you participating in the existing thread and wanting to broadcast this to some wider audience. in current fedi, you would reply + boost, perhaps dropping the mentions for other participants. the only thing missing from this flow is the embedding of the post you are replying to, but that's an implementation detail -- the Mastodon API could inline the replied-to post if it wanted to, and this could be interpreted as a signal to embed the post. clients are then free to render this above or below or truncated or in full or whatever.

a "breakout thread" is, in effect, you choosing not to participate in the existing thread, and instead creating your own thread. in current fedi, you would copy the link and paste it into a new post, perhaps adding a mention for the original author if you wanted to notify them. (there are some ux improvements that could be made here regarding signaling or detecting that any arbitrary link is in fact an activitypub object; look toward FEP-e232 Object Links as an example of how that could be federated similarly to existing Mentions and Hashtags.)

the biggest problem with twitter "quote tweets" is that they conflated these two use-cases. the particulars of this have been discussed to death above, so suffice to say here that it was a Bad Move, but of course twitter as a for-profit corporation has different design considerations than simply allowing people to communicate effectively.

i should also probably clarify that i am less worried about "loud replies" and more worried about "breakout threads", as from an ergonomic and accessibility perspective, at least the former preserves context. with that said, however, there is a way to do "breakout threads" properly, and they are not always unwelcome (although i think they should be generally discouraged and should only be used when necessary; replies are generally preferable to fostering healthy conversation since they force you to consider the person you are replying to). i am also worried about the "always notify" aspect, as i firmly believe this should be a (two-way) choice.

[Cassolotl]

Thanks!

[koyuawsmbrtn]

Can this feature be compatible with Misskey then? Would be bad to miss out on those quoted posts.

[eloquence]

Now that this is officially on the roadmap (MAS-48), I'd just like to encourage early communications about the intended behavior of the feature. For example:

• Will there be an opt-out or opt-in mechanism?
• Will folks being quoted receive notifications?
• Are there particular uses of this feature that will be discouraged in software?
• Are there particular uses of this feature that will be discouraged in community norms the Mastodon project advocates for (e.g., through the server covenant)?
• Are there circumstances under which the feature would be rolled back (i.e. is there a definition of failure)?

Note that I'm not expressing an opinion here on the above questions - wherever different folks may stand on those and similar questions, I think it would be very useful to have a high level outline of the planned change early, if at all possible.

[koyuawsmbrtn]

These are good questions. I am also on the camp that some people don't want this feature and rather have it disabled on their individual instance. How easy do we want to make it for instance admins to enable it? Do we want to hide a flag inside .env or do we want a checkbox in the admin dashboard?

[bleonard252]

Will folks being quoted receive notifications?

If it's implemented properly, absolutely. We should want it to work under the hood the same as i.e. Misskey's quote feature, which does give quote notifications. Right now, Mastodon users being quote posted on other platforms or via app workarounds have no idea this is happening unless they're specifically mentioned.

Are there circumstances under which the feature would be rolled back (i.e. is there a definition of failure)?

The fallback mechanism that Misskey uses puts the link on a line by itself at the end of the message, prefixed with "RE: ". This is how these quotes appear in Mastodon at the moment.

[sorz]

If we did do it we'd like to make it something you can opt out of, in a similar way to how we plan to allow disabling replies. It's not entirely trivial.

Opt out should be retroactive: a quote could be with good intent or an attack on me. I'm happy with people quoting my post most time but occasionally some bad guy use it to attack me and I can't know it beforehand. On birdsite, I can block who quoted me, then the quoted post will become "unavailable content" and unclickable.

[brendanjones]

Putting on my product owner hat, I thought it a good idea to separate out quote posting features and any settings to control quote posting. Now this issue can narrow scope to focus on quote post implementation. #25735

[ThisIsMissEm]

Putting on my product owner hat, I thought it a good idea to separate out quote posting features and any settings to control quote posting. Now this issue can narrow scope to focus on quote post implementation. #25735

@brendanjones Not doing them together would put many users in harms way potentially, and would become another flashpoint for the claims that Mastodon doesn't care about privacy/moderation.

Adding features without thinking about the trust & safety implications of those features would be a very reckless and bad idea. There's also historical examples of features added without a way to disable that ended up causing issues (e.g., trending hashtags & trending links)

[brendanjones]

@brendanjones Not doing them together would put many users in harms way potentially, and would become another flashpoint for the claims that Mastodon doesn't care about privacy/moderation.

Adding features without thinking about the trust & safety implications of those features would be a very reckless and bad idea. There's also historical examples of features added without a way to disable that ended up causing issues (e.g., trending hashtags & trending links)

I an no way am suggesting to do quote posts without the settings to control quote posting. The opposite, in fact: I'm saying put the safety features in place first and don't wait for official quote posting features here in Mastodon (given it's been 9 years and counting since quote posting was suggested here, yet quote posting exists elsewhere already).

[samwightt]

Opt out should be retroactive

There is no way to do this. Quote posts as they are implemented in other servers (which Mastodon will have to conform to) implement them as a link. You can delete the post in question, edit it, or change its visibility, but there is not a mechanism for retroactively editing someone else's post in Mastodon or ActivityPub. Blocking the user who quote-tweets you doesn't make the content invisible to all of the people viewing their quote tweet, it still makes it visible.

[richfelker]

There is no way to do this. ... there is not a mechanism for retroactively editing someone else's post in Mastodon or ActivityPub.

There is no need to edit somebody else's post. Whether to embed the linked post as a quote or only show it as an off-site link with high friction for interaction is a presentation time decision the software makes based on quotability markup, evaluating blocks, etc. at the time of retrieval/presentation. Not modification of the data model.

[itbeard]

Up.

[shaedrich]

Regarding the harassment issue, I've definitely seen this used on other sites to start dogpiles, but I'm not sure that the bad use of the feature outweighs the value it adds in giving context and enabling conversations. I've seen Mastodon users screenshot other users' posts to criticize and denigrate already; all it needs is a link to the original status and that behavior is essentially duplicated. And the original user never knows it's happening, unlike a quote boost where they should get a notification.

@fsnk My thoughts exactly!

1. When a toot contains a link to another toot, embed a preview of the latter in the former

A current workaround is to put https://qt.any.dev/ in front of the toot URL

should it be part of the same thread? should it be considered a reply?

@trwnh What about it being a separate tab in the thread? This can be helpful if the OP doesn't immediately act on the notification but wants to get back to it and needs an overview.

Could be something like OEmbed, then the same mechanism could be used for embedding youtube/soundcloud/GNU social statuses/anything that supports OEmbed.

Is the linked resource shown as a preview? Is the preview shown above, below, or in the middle of the post?

@Gargron @trwnh Ideally, this could be configurable. Atlassian supports four different link display options:

Mastodon could do something similar:

• Display link as is (current behavior)
• Display an open graph/twitter card/oembed link preview
• Display the original text as simple blockquote

@forenta Oooo this looks niiiiiice!

Can you quote multiple objects or fragments? Why or why not? How?

It would be really nice to have something like Genius' annotations, Medium notes (would work a bit similar to #27302) or Telegram's "Replies 2.0"

@trwnh Especially, quoting multiple toots could have some benefits if a topic is discussed in different threats and one wants to summarize this