Closed rugk closed 5 years ago
It should be renamed to "replica", I haven't had time to do it
On Wed, Jul 18, 2018, 14:11 rugk notifications@github.com wrote:
I know it's common jargo0n, but I still think the term "slave" should not be used and is not appropriate.
See https://en.wikipedia.org/wiki/Master/slave_(technology)#Terminology_concerns for details and how other (big) companies do it/name their products.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/tmate-io/tmate-slave/issues/41, or mute the thread https://github.com/notifications/unsubscribe-auth/AASIZHRxaM6vXTtXWIbIFwjL_QF2Uwglks5uH3pdgaJpZM4VVEIw .
Why not server? That's the language used elsewhere in the README:
tmate server side
tmate-slave is the server side part of tmate.io.
Seems tmate-server
would be a natural choice.
I also like tmate-server
the most.
I'd also suggest renaming it to tmate-server
. Afaict there's nothing slave-ish in the way it works, both tmate instances and plain ssh clients connect to it, so "server" seems just fine.
We are pondering uploading this to Debian, and we'd better get the name right before that.
I agree. I'm going to change the name this week.
This month, I'm going to spend time on renaming all the projects and making docker images. The infrastructure needs a rehaul.
Because tmux has a server/client architecture already, it might be confusing to reuse the name "server". (both the tmate-server and tmate-client have runs a tmux server).
On the design architecture diagram, I ended up renaming tmate-slave to "replica":
Does tmate-replica make sense?, or we should stick to tmate-server?
I'd say "server" and "replica" only make sense from a tmux perspective (the contents of terminal on the server host are replicated to the replica host). When looking at the whole setup from an end user perspective, the thing running on the target host is a client that connects to a server running elsewhere in the internet.
Why not something like "local client" and "local server" on the client vs "remote server", "remote client" etc? In such a case the "browser" would e.g. also be a "local client" (that happens to be a "HTML5 participant", as you call it.).
I think the name tmate-ssh-server
makes sense from both a user perspective, and an architecture perspective. Its role is twofold: 1) provide authentication, and confidentiality to the tmate client and 2) provide sessions to SSH clients.
If we were to only use HTTP clients, we could bypass this ssh server and go straight to the proxy (which could be renamed to tmate-websockets
)
I know it's common jargon, but I still think the term "slave" should not be used and is not appropriate.
See https://en.wikipedia.org/wiki/Master/slave_(technology)#Terminology_concerns for details and how other (big) companies do it/name their products.