People,
Is there any interested in building a very cut down, minimal “Mini” Discourse that is easily built on ANY distro (my priorities would be Fedora and Alpine) but where we could experiment with some of the chat / social features mentioned in the discussion here:
?
In my experience it’s often a bad decision to sacrifice functionality for these kind of technology related reasons.
You want to fork discourse and not run it under docker? No good can come from that.
Hmm . . this article is about FlatPak but has some general comments about containers:
Deploying apps for the Linux desktop is hard.
This article is about the usefulness of containerized application on users desktops. Discourse was never meant to run locally on your laptop along side your browser, but in a server where it’s accessible for your whole team/community.
So I dont understand how it’s relevant for Discourse 
Tell us about the use case/scenario you think this will enable.
It doesn’t involved a fork, but the first thing that came to mind was: Discourse on a Raspberry Pi | Blog.
On the other hand, launching a small VPS to test out Discourse isn’t much different, and goes a long way for experimenting. 
While I appreciate the inspiration for this, I honestly think Discourse is reasonably easy to try in a few places, Digital Ocean primary among them. There are some core requirements that make it less easy than, say, almost any PHP-based forum, but that’s a whole other issue that’s not worth getting into here. 
But I do think Discourse testing is reasonably accessible to people if they’re willing to experiment in Digital Ocean. There is a certain barrier to entry, but I can’t really imagine a good way to address it given Discourse’s core development and hosting approach (RoR, Docker, etc.).
The greater question I have is whether this gets at one of the core reasons that Discourse seems to be less “in the conversation” around modern community platforms. My feeling is the answer is no, and I think there are some basic design and functionality issues that are probably a bigger factor. But I’m curious if you feel otherwise.
It is not site that is the problem, it is building on a Linux other than Ubuntu.
But if you were building a “Mini” Discourse from scratch you would have a Green Field to experiment in.
On a small LAN quite frequently there are not many differences between the workstations and the servers . .
Do you mean you want to have the host os be something other than Ubuntu and do use the standard docker container? I don’t think that’s so hard. I
The point is that the article is about Linux desktop apps and Discourse is not a desktop app.
Well maybe not for you but I spent a while getting a Fedora Dev version going but was roadblocked with the Production version . .
However my “Mini” suggestion is for just that - starting from scratch - perhaps with Alpine - and producing distro packages not Docker images . .
I am clearly not making my point very well - as I just said in the other response: ‘However my “Mini” suggestion is for just that - starting from scratch - perhaps with Alpine - and producing distro packages not Docker images . .’
So that’s what you mean. Yes. As you have witnessed, that will be hard to create, maintain, and support. You would likely need someone working close full time to keep it up to date. There are so many pieces that you would need to track, from an intricate nginx configuration (which might actually not be one of the hard parts) to the image processing pieces and those are just the obvious pieces.
The bitnami docker image does that, though, and uses a different rails web engine, so it’s possible, but, as with bitnami, you’ll be on your own to support it.
Why do you think this is a good idea? Do you think that you could do that and reduce the system requirements in some tangible way?
Well I was not confident that my idea would get much support but I thought in the light of the previous discussion about Discourse not being considered for some situations then maybe a re-evaluation and experimentation might be useful - so I thought it was still worth a post. However, I thought such an effort could only be done by starting from scratch - BUT with input from existing Devs about what a core, minimal app might need.
I still feel a smaller, lighter and simpler app / package could be useful for smaller shops.
In the unlikely situation that others thought the idea was worth pursuing, I would help with the exercise as much as I could of course!
I think that the issue is that this is continually changing.
The docker version works on a Raspberry Pi. I’m not clear what “smaller shops” would want to target.
I indeed don’t have the slightest idea what you want to do. I thought I did but it’s getting more and more confusing.
You want to “start from scratch” but also a “cut down” version. You want to build it easily on any distro and make it simpler at the same time. And you quote an article that talks about how hard it is to deploy apps on Linux Desktop that discusses start up times and graphics drivers, which all seem pretty irrelevant to me.
For now I am sticking to my first response (which has been ignored so far)
and what @pfaffman says is very true indeed
Still this:
Discourse isn’t a desktop application, it is intended to be installed on a server and connected to by remote clients via a browser. You continue to elaborate on how you think discourse could be pared down, but haven’t given a use case as to why.
What’s the purpose of installing discourse on a desktop distro. Which scenario isn’t supported by the current installation?
Discourse can run on a $5/mo VPS or a $35 SBC over most domestic internet connections. How much smaller does it really need to be?
Are you proposing this work as a way to help address the “Discourse not being mentioned in conversation” problem (e.g. by making it more accessible for more people to try it)? Or is your intention different than that? I must say I don’t entirely understand your actual goal, separate from “make Discourse easier to setup on multiple distros”. By which I mean, I don’t understand what broader goal that would accomplish, what greater need is being served. Greater potential for experimentation?