This review would have been much better if written by a different author. This reviewer seemingly can't get over his biases for the hyperoptimized consumer market, and isn't the least bit interested in servicability or software freedom. These things are certainly trade offs and that needs to be acknowledged, but the review constantly "others" the desire for such things rather than critiquing the device within that context. This leads to useless critiques like focusing on the lack of processor speed while concentrating on bloated webapps. The author even forgoes trying "Maestral, a lightweight macOS and Linux client for Dropbox", which if truly lightweight should have run fine!

Meanwhile it sounds like the MNT has some real flaws (the battery discharging thing), but the context of that isn't really described - can the battery controller IC use much less power but the proper code hasn't been written, or is it a true design mistake, or what? In general open hardware involves some very hard tradeoffs, and this review just complains about that difficulty the whole time while pretending not to.

I appreciate that the reviewer takes more of a perspective of a mainstream consumer, and he does acknowledge that people caring about this kind of hardware have ther points. Some random reader stumbling upon this article wouldn't fare "better" with this laptop (well, they possibly could if they sought help from the community...). So it's a fair representation. If someone wants to appreciate this product, they probably have to adapt a slightly different mindset, not fully communicable by a review.

I am interested in MNT Reform for a long time myself, though haven't pulled the trigger yet. The different architecture thing (= some inconveniences with software) and the question of whether it would handle Zoom calls (with an external webcam of course) are among good points to consider. One definitely would have to plan their productive use around this laptop if it were to become a daily driver. I'm used to doing my serious business computing on a desktop, but still.

> So it's a fair representation. If someone wants to appreciate this product, they probably have to adapt a slightly different mindset, not fully communicable by a review.

Yes but the review could and should be within that mindset, while making clear that it differs from a mainstream consumer product. Instead this review is like 'I guess if you care about <mindset>, then you're fine with <subpar spec>', while never really reviewing <subpar spec> within the context of other comparable options.

Like why did they choose a Freescale quad A53, rather than say the hex-core RK3399 of the Pinebook Pro? I assume it's because the system-on-module makes it possible to upgrade. Explaining that in-mindset tradeoff would be valuable, but it's completely missing.

Furthermore there is a comparison of mindsets to be done. For example if you are concerned with servicability but not Free software or ARM, this probably isn't the machine for you.

FWIW personally I segregate my usage of modern bloatware to separate machines, rather than taking a clean trustable machine (that I'm paying for in other ways) and then polluting it with unknown blobs. Devices have become inexpensive ($1300 price tag on this notwithstanding), and it's easier to give proprietary bloatware the environment it expects rather than trying to swim upstream.

The crowdsupply page has some insight on why they chose the SoM:

https://www.crowdsupply.com/mnt/reform

> Both SoC and memory chips sit on a System-on-Module (SoM) that we source from Boundary Devices, a US company that specializes in i.MX reference designs. We selected their "Nitrogen8M SOM" module for Reform because it is the only available module for which you can download the complete schematics and understand what every component does.

> Meanwhile it sounds like the MNT has some real flaws (the battery discharging thing), but the context of that isn't really described - can the battery controller IC use much less power but the proper code hasn't been written, or is it a true design mistake, or what?

There's a known (and fixed) keyboard LPC power drain issue, which is likely what TFA described.

"fully open hardware" can be misleading: the most important parts including the CPU and motherboard controller ICs are closed.

(I know there are no performant and open source CPUs around but that's besides the point)

I'd love to see a POWER9 laptop. It'd probably be more or less the same thickness as this one.

I'd like something like this, but that uses the Pi Compute Module 4, and doesn't cost quite so much (bonus points if it can shave off half the height... the thing looks straight out of the 90s!).

This exists: (I've got one) https://liliputing.com/2021/06/devterm-is-an-open-source-por...

That is stinkin' adorable. Though I'd prefer one with a Cambridge Z88-type keyboard...

How is it? Can you touch type with the keyboard or is it too small?

This might soon be a reality: https://twitter.com/mntmn/status/1479480351651155971?s=20&t=...

Hopefully, we'll see at some point a standardized RISC-V compute module, that can be used as a building brick for all kind of devices. Besides laptops, an open-hardware server with 32 of those modules for instance would be a thing.

There are multiple modules compatible with the RPi CM3 (such as the ones used for the Devterm) and CM4 form factors. Those already can be used this way.

Since these little boards are underpowered, it'd be interesting to have a board with multiple network-connected boards (like the Pi and Pine64 cluster boards) so you could offload work to other nodes in your laptop cluster.

Regardless of other projects like this, I've been looking for a fun project for my CM4. I think you just made my weekend!

Why do you prefer the Broadcom chip to the NXP?

A72 cores are faster in this specific instance, but more generally there's a ton more support for the Pi-based boards so doing any work with them is easier than other SoCs.

That is not true. It still requires proprietary binary blobs.

Is 'requires' the correct word when the blobs are optional?

They are not optional if you want your system to boot.

Whoops, forgot about the DDR PHY one. I was just thinking about the optional blob to get HDMI working. My bad!

I hope more 3D models for printing and CNCing come out into open source for different type of computer/laptop cases. I can get all the electronics working with the specs that I want but I am absolute garbage at 3D industrial design (or any design for that matter).

Interesting that they say the kailh choc brown switches offer clickiness. I have a keyboard with kailh choc browns and they feel like they have zero clickiness to me. Remind me of full size red switches.

I have the MNT keyboard on my desk at home and its quite nice. I dont think it's clicky like a model M but it certainly makes a lot of noise. A fellow hacker who owns a Reform said they were notified of the excessive noise both at home and at a local coffee shop.

My only gripe is the keys feel a little cramped and I had to get used to the weird del key placement and the split space bar. Though I still much prefer the happy hacking 2, the MNT keyboard on a laptop sounds very nice.

I have a Reform and totally agree. I only had prior experience with Cherry MX blue & brown, so I expected more tactility from what Kailh is calling choc brown.

MNT recently announced standalone keyboards with other Kailh switches as options. It's using the same PCB, so I've ordered one with choc white and plan to do a board swap with my Reform.

Slower than a Pi, and it's not even close... ouch

The Reform uses an upgradable SoM, there's already an RPi CM4 based unit in the works.

What's great is you'd keep the user-serviceable batteries, nice keyboard, modern display...