installer … with enhanced web UI
∘ ∘ ∘ ( °ヮ° ) ?
seems like yet another electron app that only runs locally. i’m guessing that hiring traditional desktop ui developers is getting harder and more expensive over time, so they don’t bother anymore and just hire webdevs instead
The purpose of having a web-based user interface is to enable remote installation using a web-browser, which I imagine could be a pretty common use-case for server installation. Since a web-based interface is necessary for remote installation, it makes sense to use the same interface locally instead of having two independent interfaces.
Makes sense, especially for server environments
A quick glance at the Agama repository suggests that the server is written in rust and the front end in react. I’ve no idea how it all works in practice as I don’t use Tumbleweed any more. I really liked the yast installer but it was getting old.
It worked perfectly well to install though? Like why does it matter if its old
That’s kind of like asking why we’re not all driving Ford Model T cars, after all you could drive in them just fine. Technology moves on, best practice moves on, Hell, everything moves on.
anyone who thinks web tech is best practice on the fucking desktop should be expelled from the whole field
said ford would consume way too much gas and produce way too much noise, among other things. but what’s the problem with the current installer? that it doesn’t have curly corners, and that it has too many options which is confusing to those with no reading comprehension?
So I suppose you never use a browser to run a web application on the desktop :thinking_face: Anyway it;s a client server architecture designed for remote installation on servers as well as local installations. It makes sense to have one installer do both.
As to the old installer, when you knew about the un-obvious features, it was brilliant from a user perspective, but I’m willing to bet that from a developer perspective, it was hard to maintain, hard to add new features to, and fragile.
I’m already running too many browsers in some trenchcoat on the desktop.
Anyway it;s a client server architecture
that does not warrant a browser. not only JS can do HTTP requests
It certainly worked and was full featured, but the interface wasn’t very good. Having to edit the network interfaces to configure them wasn’t good UI for example (the partition editor works the same way). It also took until my second install (that was quite some time ago) to figure out that I could pick what software I wanted to install.
Anyway, a lot of things could be made clearer for first time users.
I’m not a SuSE user - but did they not have the ability of using LVM or setting the hostname during install (GUI) prior to this?
I don’t know about LVM but hostname was configurable in the old/current installer, this is news for a new installer to replace the old one.
That’s great ! Is there any chance we will switch to Agama instead of Calamares? Cause instead of fixing the LVM installation issue they just drop the support…
Great if you are allergic to using anything that dosent adhere to the latest design style, I guess
But the last installer was more than enough.
I was considering switching from Windows to Linux but man, I don’t know, if only they had one more installer… /s
Why people can’t just use Calamares and stop reinventing the wheel?
Why people can’t just use <name> and stop reinventing the wheel?
This is a really dumb take. If you don’t like it, don’t use it. They got an idea and implement it? Good for them And what does that even have to do with switching to Linux?
i’d usually agree, but in this case, it feels like a cost-cutting measure. webdevs are cheaper and more available, so it’s cheaper for them to just rewrite the installer in electron than pay more expensive desktop developers to maintain their existing installer
Assumptions are assumptions. The server is written in Rust, the idea is to be flexible with control and the optional UI. It has a big focus on Enterprise and things that were difficult with YAST are easier with Agama, such as unattended installation and using Ansible. For a simpler use case you can boot it up on your headless server and connect to https://agama.local/ in a web browser and continue the installation.
-
Instead of wasting time on something that has already 1k iterations, they could redirect that effort on something beneficial to Linux.
-
Any person contributing to this new installer is a person less contributing to something distro agnostic, which is a loss for everyone.
Why isn’t everyone working on Arch instead of wasting time with all those other distros?
Never used arch but I share the same feeling for distro sometimes. A linux distro can either:
- Introduce a new way of doing things, or some big novelty
- Being a complete waste of time and effort The vast majority is the second case, but sometimes we are lucky and we get something new.
-
Because they don’t fix a 5 years old bug about LVMs? instead they write a post on how to drop support for it?
Right now you have to do some sketchy bug/workaround to make it work with LVM…
So yeah drop calamares, hello new installer :)
sure making a new installer from scratch, whats more with web tech, is definitely easier than fixing the existing solution thats almost perfect
