Often a flag worthy post would contain flag worthy content mixed with acceptable content. Additionally there may be multiple types of flag worthy content within the same post. It would be advantageous for both better educating the poster of the flagged post and for decisions of agreeing/disagreeing with flags if while flagging you could highlight multiple parts of the post you are flagging (much like how quoting in Discourse works), and indicate the specific reasons each part was subflagged. The subflags would all still be part of a single flag for that post. Each subflag should also allow agreeing/disagreeing to aid in the decision process.
3 Likes
Good suggestion but we probably need a mockup of what this might look like. When flagging a post - you select from a (narrow) set of reasons(â˘) for the flag, and for âsomething elseâ, you can send a PM. The problem(â˘) is that the PM doesnât use a standard post text input, but instead just shows on the flag options window.
What would be nice is to instead open a full post input editor window for the pm, showing the full contents of the post above that, and in addition to the âquote replyâ button for selected text, show âflag textâ which would make add it to the post editor and make it appear similar to a quote (but maybe have a different background like (the warning color probably)). For each âflagâ, in addition to editing author (like quote), set a âreasonâ too - or just type it out as part of your reply (below the flagged text display in your post editor)).
4 Likes
Maybe a way to compliment the existing system could work nicely, something like:
Step 1) - A user clicks the Flag icon on a post.
Step 2) - User picks âSomething elseâ from flag options window and is shown option to choose to either:
A). Send a PM directly to the post author.
or
B). Reply in-line to the post with a âprivate replyâ sort of like âwhisperâ text (your reply will only be visible to you, and the flagged post author, and mods and admins).
For either choice, use the full post reply text editor (and never do this in the flag options window). Alter what info and icons are shown at the top.
If selecting (A) to message the person you are flagging, directly, show â@thatUserâ pre-filled and add any additional @users or @groups here.
Change icons next to the reply-arrow icons and âTopic titleâ display to indicate:
with A) show â
The topic titleâ
or B) show â
The topic titleâ
Step 3)
In the case of A, quote anything you want from the displayed flag post and then send the message as a private direct message.
In the case of B, treat the reply like a whisper text(?), and show a new option âFlag textâ when selecting text from within the flagged post. Example:
This would add in a new (type of) block, formatted almost exactly like quote text, but with optional editable reason:
[flag="romdos", post:2, topic:43040", reason="Why this being flagged"]
Some problem(â˘) text being flagged[/flag]
Which would display in the private reply (and post preview window) as:
[quote=âromdos /then right-align the reason field -->/ ââWhy this is being flaggedâââ , post:2, topic:43040"]
Some problem(â˘) text being flagged.
(Pretend this has the warning color background)[/quote]
âŚwith the âreasonâ field displayed directly on the the flagged-text block.
Step 4) User types out any additional comments, and clicks âReplyâ in the post editor.
Once posted, admins & mods see:
âAgree 
0â & âDisagree 
0â links next to each flagged-text blockâs âreasonâ (or below the block), each with click counter.
The most significant impacts of this proposed feature, would be:
On the admin / mods flags screen ( /admin/flags/active ), the full content of the post is displayed for the admin/mod to inspect, and whatever text got âflag textâ'ed (by any/all users) would appear as specially highlighted/sized/colored/special warning color background to easily draw their attention to specific parts of the content.
And finally, instead of just showing the 1 reason for the flag (currently like âinappropriateâ or âoff-topicâ), show all reasons, each with the number of âAgreeâ'ed and âDisagreeâ'ed, which would be useful for an at-a-glance overview of other modâs opinions - before you might âpull the triggerâ by using this screenâs âAgreeâ, âDisagreeâ, âDeferâ and âDeleteâ buttons.
2 Likes
I have the same deep objection to this as I do to annotation systems. And I basically hate annotation systems, so fair warning. 
We already have so few flags; making the process even tweakier and more convoluted isnât going to achieve anything useful.
3 Likes
We have so few members flagging anything other than obvious Spam that I started a poll/topic to try to find out why.
One of the replies, from a long-standing, useful member who does flag stuff read:
Well, âspamâ is kind of a catch-all, isnât it? Self-promotion is spam. Fluff posts, such as if someone copy-pastes Wikipedia to pad their post count, is typically spam. Off-topic posts that are so severally off-topic to warrant a flag are almost always spam. That only leaves trolling, which I would flag as inappropriate, except I donât see those very often.
If the current system is already so difficult to get to grips with as that post suggests, Iâd be very much opposed to the idea of making it any more complex. I think people would just give up in despair.
1 Like
Existing discussion on the Flagging UX for the dialog here:
https://meta.discourse.org/t/some-ideas-for-improving-the-flagging-dialog/22072?u=deanmarktaylor
1 Like
To be fair, the current flagging system is heavily under-utilized on my site. Users donât seem to understand that flagging a post is the desired method for them to help to âmoderateâ the forums.
Itâd be nice to be able to edit the flag reasons (globally) to in/exclude specific reasons or include category-specific reasons (a category for research might have âcitation neededâ).
Additionally, the post editor itself on the flag options window is terrible. Iâd highly prefer to use the full editor, and never have to PM a user about their post from the options screen (where I canât even see their post because the options window is covering it). If I could quote specific parts of their text to discuss it with them, insert links easily, and not be limited to 500 characters (or whatever it is), thatâd be a big improvement.
The system I proposed would expand on already-existing features, making use of âwhisperâ-like hidden-post to discuss the flagged post in the same place as where the flagged post resides.
EDITS: Basically some of this is expanding on the ideas from the âCreate custom flagsâ thread, which I just read.
I like some of the concepts/mockups shown in the âSome ideas for improving the flagging dialogâ, but all these mockups still show using the options window (directly) to send messages or leave ânotesâ, which is bad, for several reasons (IMHO).
It lacks a full text input editor, and also because the âsomething elseâ PM moves discussion of a post away from the post itself (to messages/inbox), when you in fact want the person receiving the flag to be where his flagged post is, for him to edit it immediately (which is fixed by having a âwhisperâ style private-reply appear inside of the thread).
1 Like
The general idea of this proposed feature is to make the flagging system serve more of an educational and guiding purpose to both the poster being flagged and to the community member that is submitting the flag. I think that being able to better communicate and understand the specific reasons a post might not be acceptable would be more effective than having the system effectively just give a blanket ânoâ to the whole post. Only saying ânoâ to a post as a whole can be misinterpreted as saying ânoâ to acceptable/salvageable content in a flagged post and not just the problematic aspects.
Having more guidance intuitively built into the system would especially be useful given that two of the main objectives of Discourseâs flagging system (as I understand it) are to encourage the flagged posters to edit/move their flagged posts to become acceptable and useful (if the post is salvageable), and for the community of a given forum to be generally as self-corrective as possible with minimal moderator intervention required.
1 Like