Compare commits
1 commit
main
...
post-mm-th
Author | SHA1 | Date | |
---|---|---|---|
e9f07f105e |
1 changed files with 72 additions and 0 deletions
72
content/posts/im-threads.md
Normal file
72
content/posts/im-threads.md
Normal file
|
@ -0,0 +1,72 @@
|
||||||
|
---
|
||||||
|
title: "IM Threads"
|
||||||
|
date: 2018-09-22T22:00:00+01:00
|
||||||
|
draft: true
|
||||||
|
---
|
||||||
|
|
||||||
|
I have had the opportunity to use [Mattermost](https://www.mattermost.org) for
|
||||||
|
over a year, and I thought I would give some feedback on a feature I still
|
||||||
|
have troubles dealing with: **threads**.
|
||||||
|
|
||||||
|
A bit of a disclaimer, Mattermost is the only solution of its kind
|
||||||
|
(centralized, non-standard, one blessed client) that I have used extensively.
|
||||||
|
I do not use Slack, Gitter, Discord or what have you. I generally use XMPP and
|
||||||
|
IRC (via XMPP), in my terminal.
|
||||||
|
|
||||||
|
I haven't gone through all the stages of grief yet concerning this topic, so
|
||||||
|
please bear with me. This article addresses usability issues with the web
|
||||||
|
client. I do not have solutions to offer, but I hope that my criticism will be
|
||||||
|
taken into account, and I would be happy to work with the project to try and
|
||||||
|
improve the feature.
|
||||||
|
|
||||||
|
## What I understand people want
|
||||||
|
|
||||||
|
- A way to group discussions by topic.
|
||||||
|
- Being able to talk without having to state context all the time.
|
||||||
|
- Being able to refer to these discussions in their integrity, (via some URI,
|
||||||
|
or else), to even read days later.
|
||||||
|
- Not having to highlight every concerned party in every message.
|
||||||
|
- Not being highlighted if you are not part of it
|
||||||
|
|
||||||
|
Out of context this feels like people just wanted _channels_, because this is
|
||||||
|
exactly what I described, but certainly details are important here. What I
|
||||||
|
think people want is _fine grained_ all-of-the-above, that is, have usual
|
||||||
|
channels set a common topic, e.g., "programming", and be able to split
|
||||||
|
discussions happening in this channel even more, with some kind of nested
|
||||||
|
channels, or threads.
|
||||||
|
|
||||||
|
---
|
||||||
|
WIP
|
||||||
|
|
||||||
|
This last bullet point is actually the most important part.
|
||||||
|
|
||||||
|
## Implementation case study
|
||||||
|
|
||||||
|
Mattermost, as other IM solutions, regroup most of these features under what
|
||||||
|
they call __threads__.
|
||||||
|
|
||||||
|
I'll try to report a couple of things to upstream if possible, in the
|
||||||
|
meantime here's what I think about it:
|
||||||
|
|
||||||
|
- In MM: missing fork feature: after 3 messages in a thread it's already
|
||||||
|
diverged from the original discussion, and might get back to it at some
|
||||||
|
point later on. What changes from a normal channel?
|
||||||
|
|
||||||
|
- In MM, UI is unintuitive:
|
||||||
|
* Not easily accessible. You have to move your pointer to the small icon at
|
||||||
|
the far right of your screen.
|
||||||
|
* No keybindings to navigate between threads, that makes them a hard to use.
|
||||||
|
* Thread messages appear in the channel, that makes the whole thing
|
||||||
|
unreadable when there are multiple threads active at the same time.
|
||||||
|
* Bug? threads window not scrolling down automatically.
|
||||||
|
* possible resolution: change into short-lived channels?
|
||||||
|
* Can only have one thread opened at a time. That means if I want to follow
|
||||||
|
a thread, I can't reply to/follow any other threads.
|
||||||
|
|
||||||
|
- Too much micro-management to get right
|
||||||
|
|
||||||
|
|
||||||
|
- Mattermost should allow users to attach anything (Direct Messages, Channels)
|
||||||
|
where the current thread UI lives, so they don't have to start a "fake thread"
|
||||||
|
just to be able to counter the limitations of the tool. Especially in 1:1
|
||||||
|
Direct Messages, where most of the benefits of threads are moot.
|
Loading…
Reference in a new issue