digression :)

This commit is contained in:
Yann Esposito (Yogsototh) 2019-10-16 14:26:26 +02:00
parent 1f7d79ba06
commit ae4251f2d9
Signed by untrusted user who does not match committer: yogsototh
GPG key ID: 7B19A4C650D59646

View file

@ -15,8 +15,8 @@ is IRC + ZNC (with replay module) + weechat (with replay script) +
thelounge + Palaver (on iOS).
The two game changer are:
- znc replay module (history handled correctly)
- znc palaver module (push notifications on iOS)
- ZNC replay module (history handled correctly)
- ZNC palaver module (push notifications on iOS)
#+end_notes
How to chat in 2019? Certainly with slack, or via a social media app in the
@ -33,7 +33,7 @@ Here are the feature I think a modern geeky solution should have:
a lot, you will prefer density over good looking.
Most app, web app are terrible related to information by number of pixel
ratio.
2. *multiplatform*: If you do not have a terminal at hand (or emacs) then,
2. *multi-platform*: If you do not have a terminal at hand (or emacs) then,
you should be able to get your message on your phone or via a web
interface for portability.
3. *self-hosted*: you should control your data, your history, your logs,
@ -41,7 +41,7 @@ Here are the feature I think a modern geeky solution should have:
4. *teams* and *direct messages*
5. *notifications*, I tend to control those a lot, but a small private team
chat is one of few exception where you generally want to be notified.
6. *Frugal*. Really, we have a responsibilty to do our best not to consume
6. *Frugal*. Really, we have a responsibility to do our best not to consume
more resources than we really need.
Chat should be about TEXT, not images, not videos, not presentations and
PDF.
@ -51,26 +51,65 @@ Here are the feature I think a modern geeky solution should have:
taking the time to answer correctly.
8. Use free software (not just open source software).
I am quite disapointed by /modern/ chat applications.
I am quite disappointed by /modern/ chat applications.
Their major problems are:
- opposite of frugal (slack consume 400Mo of RAM for me, the technology is
an electron app, or an equivalent complex web app consuming the same
amount of RAM)
- prodigal; most those client applications (slack, gitter, riot,
mattermost, etc...) easily consume more than 300MB of RAM.
Most of the time those clients are all electron app.
- not private. Most solution do not encrypt your conversations. Even
if using encryption mechanism and you trust your client, and you will
still reveal your social network topology.
- opposite of minimalist and solution focused, I want dense, ascii text. I
do not want, emojis, images, gifs, videos, messages that will improve my
"engagement" with their app, feature that increase the social drug used
by most social network platforms. Typically, show when someone is typing,
show when you read a message, get notifications about missed messages,
etc... For me those are anti-features.
- anti-minimalist, I want dense /text/.
I do not want:
+ emojis,
+ images,
+ animations (gif or videos),
+ HTML/Markdown display
- Manipulative; they try very hard to optimize engagement.
This is generally achieved through FOMO[fn:FOMO] and social anxiety
manipulations.
A few examples:
+ show when someone is writing a message
+ show when someone has read a message,
+ get notified about missed messages,
+ get a "top messages you missed",
+ etc...
- Prepared for EEE[fn:eee].
Most of those "modern" solution are a all-in-one solution, server +
clients with generic + specific features.
That way, some feature will be only possible in their client or that will
push concurrent app developers to follow the pace of new features.
** Different tries
The /manipulative/ part is quite important in fact.
Many of those "features" are in fact wanted by the users.
I think this is because those people should take the time to reflect about
the real plus value of those features vs their price.
Like it is state in Digital Minimalism[fn:dm].
Even recent projects like matrix, mastodon that thrive to break monopolies
are going a huge speed into this eternal run software development new
ideology.
Always evolve by adding as much new features as possible, instead of
striving for simplicity, stability and security.
But do we really need those features?
People are asking for them, but we pass too much time asking about the how
and forgetting about the why?
Each new feature should be really studied before being developed.
Would that feature be a full total net positive for everyone?
Or is there a price to pay? The price is generally the size and complexity
of the project, that will make it bigger, slower, harder to manipulate,
install, interact-with.
[fn:eee] [[https://en.wikipedia.org/wiki/Embrace,_extend,_and_extinguish][Embrace, extend, and extinguish]].
[fn:dm] http://www.calnewport.com/books/digital-minimalism/
** About failed attempts :PROPERTIES: :CUSTOM_ID: different-tries
:PROPERTIES:
:CUSTOM_ID: different-tries
:CUSTOM_ID: about-failed-attempts--properties---custom-id--different-tries
:END:
:END:
- Matrix: I've used Matrix, and in fact it was really good except; the
@ -101,8 +140,9 @@ It was the easiest and best solution.
5. it is minimalist
6. it is feature complete for a small team chat.
[fn:FOMO] Fear Of Missing Out
[fn:tries] Here is a list of the chatting solutions I used for some time
and finally abandonned (I certainly forgot a few ones):
and finally abandoned (I certainly forgot a few ones):
- slack
- matrix (self-hosted)
- keybase