Comment by zeta0134 | Hacker News Reader

Comment by zeta0134 | Hacker News Reader

Parent Full thread

You would think this, but it's preeeetty amazing how much of the Linux world works on an ARM-based Raspberry Pi. It's mostly a matter of finding the source and doing a proper cross-compile, but it's kinda incredible how much work the community has put into making the little board usable. The later models and their substantial bump in available RAM have made desktop computing plausible: I can run Blender to produce a 3D model, switch to Utilimaker Cura to slice it, then run Pronsole and send the .gcode out to the printer, all on the same device! It's not particularly fast, but it's usable at least, and the driver support an OpenGL in particular have come a long, long way.

If a $35 (okay, $55 for the ones with 4 GB RAM) computer running an older down-clocked ARM can do this? I have little doubt that a POWER architecture could theoretically do it just as well, if not better since it should presumably be a much faster chip. I think it's mostly going to come down to support from both the vendor and the community, but I wouldn't write it off entirely.

> I have little doubt that a POWER architecture could theoretically do it just as well, if not better since it should presumably be a much faster chip. I think it's mostly going to come down to support from both the vendor and the community, but I wouldn't write it off entirely.

I actually think it'll come down to the validity of that first sentence I quoted. A Pi 4 definitely has relatively decent specs for a small server or low end laptop.

If this isn't a Power9 (aka the series that the G5 and the latest Power arch chips are a part of) then all this has going for it is the openness of it. Is it more open than a Pi 4? If they release a successor to the Compute Module 3+ with better specs and someone builds a laptop chassis around it, is there any reason I shouldn't just buy that?

The G5 (PowerPC 970) was based on POWER4.

And a POWER9 is has quite a bit more horsepower than a Pi.

Are you saying the end user should do the cross compile, or the package maintainer? I feel like for arm it's mostly been package maintainers who've been increasing the available software. For PowerPC, if the user needs to compile from source that is a big reason to avoid PPC IMO. A bit of a chicken and egg situation.

For the Pi in particular it's been a mix of both, I think. Its sheer popularity means that there's less pressure on the package maintainers to ship working builds; for example, you can technically "apt-get install cura" but it won't run, it crashes during initialization for unclear reasons. I'm running an AppImage of it put together by a community member. It's both exciting that this is even possible, and frustrating that it took so much research and tinkering.

I'm not sure why you're getting downvoted, because this is a really legit angle to argue. Philosophies aside, I think the out of box experience is critical to adoption of a device like this, and it's a big risk for Slimbook to take when switching to a much less popular architecture. For the Pi it was acceptable because it was sold as a development board for education, so "batteries not included" is a feature. For a device poised as a workstation, I'm not sure how well the frustrations of early adoption will go.

It's exciting that Slimbook is willing to take that risk, but it'll be up to the community to determine whether the devices (and their vendor support) are worth the cost of maintaining the packages in the long run.

EDIT: For clarity, the out of box experience on the Pi here refers to Raspian, which is a debian-based image put together by the vendor. There are lots of other distributions that can run on it, and some of them might have better luck with the edge cases I hit.

Reply on news.ycombinator.com