Page MenuHomePhabricator

Adjust the indentation level for direct mentions
Open, NormalPublic

Description

The new indentation logic makes nesting more meaningful by encouraging linear conversations (until a tangent conversations appear). Although the norm when replying to the last item in a conversation thread is to continue the conversation, the new model does not provide a way to start a tangent conversation intentionally to those users that just want to make an off-topic comment to the last item.

One way to support that is to consider as tangents those comments where the author of the previous comment is explicitly replied to (i.e., the reply starts with a mention to that user). That is, If I'm at the initial indentation level (level 0) and I reply to the last comment by Bob starting by "@Bob..." my intent is likely to make a side comment and the system can support it by posting it to a new indentation level (level 1).

We can apply this logic to the first indentation level (where mixing tangents and main conversation can have more impact in terms of noise) or to all levels (where there is a risk of the approach being abused and leading to "diagonal conversations").

Below, I created an example conversation which illustrates how the proposal may help to better structure a conversation in some specific cases:

Example Conversation: "Who invented the telephone?"

  • Anne: "It was Bell"
  • Bob: "It was Yuri Gagarin"
    • Anne: "@Bob we are talking about first time sending voice through a wire, not people to space."

With the current model, the last comment would be at the initial indentation level, which is not terrible, but may get in the way of someone just scanning the initial indentation level to check the proposed answers to the initial question.

See T105438: Move indentation model to rendering

Event Timeline

Pginer-WMF raised the priority of this task from to Needs Triage.
Pginer-WMF updated the task description. (Show Details)
Pginer-WMF added a subscriber: Pginer-WMF.
Restricted Application added a subscriber: Aklapper. · View Herald TranscriptMar 30 2015, 7:34 AM
DannyH triaged this task as Normal priority.Mar 30 2015, 5:19 PM
DannyH added a subscriber: DannyH.

I want to see how people use the new model before we make more changes. Have you seen examples where the new indentation model is confusing?

I want to see how people use the new model before we make more changes.

This is just to capture the idea on how to provide a way to start a tangent intentionally in a non-prominent way.
I totally agree to wait until we observe the model working in practice, to identify if the need exists and in which context.

Have you seen examples where the new indentation model is confusing?

Nothing clearly. Some comments such as this one identify a conversation as sub-optimally structured. However, that conversation mixes real conversation with meta-conversation (about the indentation model itself), and it was created when the "reply link" was hidden for the last topic, so it may not constitute a valid real-world scenario.

qdinar added a subscriber: qdinar.Feb 6 2016, 7:52 AM

the idea is ok ... but if they do so, it will be impossible to [automatically] reshape/reconfigure/modify discussion to classic diagonal threads , which/how is requested in T93024.

DannyH set Security to None.Feb 8 2016, 7:23 PM
DannyH removed a subscriber: DannyH.

Another possible approach is to always provide options to do both, continue the conversation and create new tangents, but encourage the former by default to avoid discussions to go diagonally for no purpose.

In the example you can see that the main reply actions are for continuing the conversation, but an "add a side comment" is always available to start a tangent conversation about a specific comment.

Qgil added a subscriber: Qgil.Sep 8 2017, 9:34 AM

I want to see how people use the new model before we make more changes.

I have been used the model almost on a daily basis for years, and I can assure that @Pginer-WMF has a point here. I am totally cool with the structured conversations and the indentation model restricting my flexibility for the common good. However, sometimes I do want to force an indentation level, and that is most of the times when I am replying @someone.

Have you seen examples where the new indentation model is confusing?

I can recall two types of frequent examples:

  1. One conversation is going on a tangent, yet it is "populating" the default level (no indentation). This makes it look less of a tangent and in a way makes it more relevant than it is to new readers and potential contributors of that topic. See for instance https://www.mediawiki.org/wiki/Topic:Txn15jri131au12s -- I would have preferred to force an indentation in my first reply to IKhitron, leaving the main track for new threads.
  2. Two users have left comments at a same indentation level, one after the other. I reply to the first one, and an additional indentation level comes with it automatically (good). Then I reply to the second one, which is the last post of the thread. I would like to force an indentation level to make my both replies symmetrical. However, the second reply will be land at the same level because is the last one, and this visually brings a feel that my first reply was more tangent/secondary, while what really matters is the second reply (because it is at a more prominent level). I don't have a public URL at hand right now, but it happens frequently.
Qgil awarded a token.Sep 8 2017, 9:35 AM