Please drop Fedora IoT from "Fedora Linux is available pre-installed..." Objective

Split from Objective Review: Fedora Linux is available pre-installed on more systems from more vendors - Fedora Discussion

Please drop this statement. It’s too specific and not helpful, since it’s exclusionary for non-IoT use-cases on alternative architectures.

Would you feel better if the sentence explicitly said “IoT and embedded systems” or something to that effect? The statement doens’t say only include Fedora IoT, but gives a specific target. Clearly we won’t be upset if these devices ship with, e.g. Fedora Workstation. It’s reasonable to give a specific area for growth in a strategy document.

1 Like

I don’t think it’s reasonable to constrain in this manner. The example used is the Raspberry Pi, which has desktop-type form factors as well (Raspberry Pi 400). Single board computers almost universally come with the required hardware to behave like a desktop or laptop, because the original goal of the Raspberry Pi was to democratize exploratory home computing.

The Objective’s description does not mention a Fedora Linux variant anywhere except in that one statement. That means for the success of this objective, the variant shipped doesn’t actually matter.

1 Like

It’s not meant to be a constraint at all. To the contrary, the paragraph that’s from begins “This extends beyond desktop and laptop hardware…”, and that’s simply an example that follows. Rather than dropping it, give me more examples and I’ll add them.

But also, please don’t get too hung up on the explanatory / background text. That’s easily changed and updated.

Why do you want to mention Fedora IoT so badly?

Oh, that’s easy! I’ve been talking about this for a long time — and it’s specifically why I’ve pushed for having IoT as a Fedora Edition. Editions are designed to address a distinct, relevant, and broad use-case or userbase that a Fedora Edition is not currently serving[1]. I believe IoT is a particularly important use case for bringing students and non-IT-background hobbyists into open source (and computing in general, really).

I’ve told this story before: I got excited about computers because my elementary school classroom had them available — two beautiful Apple ][ computers[2] in the hallway outside every classroom, with a small selection of software and a stack of “Teach yourself BASIC” books. I could play Moon Lander or Oregon Trail, and those were fun, but the programming books felt like I could make them do anything. And, basically, I could: there really was very little in the most sophisticated computer game available that one clever person couldn’t replicate.

And this was true not just with Apple, and for a number of years. Games like Commander Keen and Duke Nukem (the original 2D game, not the horrific 3D mess) were basically one-person creations. A lot of time and work, sure, but, you could do that. Even the original Doom — a five-person team.

Fast forward to me trying to teach Scratch in an after-school program when my daughter was in 3rd grade. Way more powerful than the systems I worked with — and a lot easier to do cool things quickly. And a few of the kids had fun … but the overall reaction was “yeah, okay, I made a line move — now, I want to make something like Fortnite. Tell me how!”

Software has come so far and become so complex that the baseline for doing something interesting is off the charts. Computer games routinely have production budgets exceeding all but the most blockbustery movies. Anyone who has that feeling of “I, by myself, can make this computer do anything!” will soon hit reality — and maybe just give up.

But — IoT is different! You can set up Home Assistant or ESPHome, and build useful little sensors and devices programmed with CircuitPython or Arduino that do actual cool useful stuff, and actually can do it better than off-the-shelf home automation. Or you can run a model railroad. Or make a clock that tells where your family is rather than the time. Or so many other things. The world of possibility is there again.

We need that, and we need to tap into it and connect to Fedora. In fact, this is important enough that I’m open to making it a stand-alone Objective next to one more focused on laptops and desktops.

  1. from the policy ↩︎

  2. in later years, fancy IIc systems with 80 character displays and lower-case letters! ↩︎


That’s all well and good, but none of that matters from writing the Objective. In some respect, it’s an implementation detail of how to broaden the adoption.

And also, Fedora IoT Edition isn’t necessarily the platform for that, since it’s not designed to help you do things. Raspberry Pi OS (fka Raspbian) goes out of its way to provide an approachable experience to do that stuff. Fedora IoT Edition does nothing of the sort because it’s basically the upstream for RHEL Edge. There’s no alignment.

So this goes back to me saying that mentioning a specific variant is too much here.

The Objective is (in current form): Fedora Linux is available pre-installed on more systems from more vendors.

As I said, please don’t get too hung up on the explanatory / background text.

This isn’t really the place for this kind of complaining. For what the IoT Edition is designed for, I refer you to IoT Product Requirement Document :: Fedora Docs. The scenario I am describing is included, so I’m satisfied at a design level.

That is completely uncalled for, considering you’re using all that as a justification for specifically mentioning Fedora IoT in this Objective.

Make a separate objective about Fedora IoT if you want that. But Fedora being preinstalled on more systems as a general objective should not be variant specific.

My point is that using Fedora IoT in the text of the Objective sets the tone of how you expect the Objective to be accomplished. It effectively creates a constraint of vision, if not of practice. Your intent might not be to do that, but your words are.

I’ve split this out into a new topic so the main one isn’t overwhelmed by the two of us going back and forth on this detail of the explanatory background text.

Neal, I asked for additional examples, which I think should resolve your concern about this appearing to be a constraint.

Most of the lab variants would qualify for this. Python Classroom and Robotics labs come to mind. Even the generic Fedora KDE and Workstation variants are valuable here.