Trying to build the ultimate Raspberry Pi computer – Zero Terminal V3 | Hacker News Reader

Trying to build the ultimate Raspberry Pi computer – Zero Terminal V3

n-o-d-e.net

<grumble>HN has given N O D E love for the design aesthetic of the presentation and video format recently, but it feels like I'm being fed the details via Morse code. It's infuriating! There is no summary, no overview, no block diagram, just a wall of transcribed prose and gamble on a non-indexed, un-subtitled 8 minute video. I imagine many people (myself included!) will just go "Okay, there a black brick and screen thing with an RPi zero (or is it a different model?) in it ... or something" and move on.

Perhaps this is intentional in the same way the OpenBSD homepage is purposefully ugly, to scare off people and create a small effort-moat.</grumble>

> purposefully ugly

I've been using OpenBSD for about 20 years so I feel like I can talk about this issue.

If you browse web.archive.org for old copies of the www.openbsd.org site, you'll notice that the current design is from many years ago, back when lots of websites looked like that.

It appears that the current design is from around the year 2000:

https://web.archive.org/web/20010118233800/http://www.openbs...

And that one of the oldest archived versions of the page, from 1997, already had much of the same information and organization, just with less color:

https://web.archive.org/web/19970327004719/http://www.openbs...

Over the years there have been attempts by users to redesign the site. Each had their own problems. I think that for a site redesign to be accepted, it would need to be a major improvement both technically and aesthetically, and for maintenance of the new site to be real easy. OpenBSD devs don't want a bunch of churny work hoisted on them.

Just as an aside, I've noticed in this thread that you're making low-effort, kind of rude posts about an important project that's not even related to the actual subject, which is Raspberry Pi. Why are you doing that?

> I've been using OpenBSD for about 20 years so I feel like I can talk about this issue. [...] you'll notice that the current design is from many years ago, back when lots of websites looked like that

If you view the source for www.openbsd.org, it reveals that it's not a matter of the site being a holdover from the time it was made, as so many replies are saying. It uses modern markup, such as header, nav, article, and footer.

OP was referring to a conscious philosophy that has been espoused by OpenBSD folks. See https://news.ycombinator.com/item?id=9240906

> I've noticed in this thread that you're making low-effort, kind of rude posts about an important project

I'm sorry if it came across like that. I didn't put this comment there to upset you or anyone, but this happens. I was referring to what appeared to be a design choices in how some projects present themselves to the world, not as a judgement around how important they are. I watched the whole video (with some skipping) for this post. My point was many people might not given the design choices made in how it was presented. I guess "important" projects don't even need to be popular or accessible to do what they need or want to do, and I was simply musing. e.g. You have an explanation for why OpenBSD presents itself they way it does because you are convinced of its value, so you see something different when you look at the homepage vs how ... nearly all? ... other people might see it. My point was is looks like it was designed by engineers with more important matters to think about, but that's some encoding not available to everyone. (And maybe that's okay, hence a small moat around the project.)

I have realised since writing this that the topic is actually "gatekeeping". It can feel icky advocating for it. OP is a talented and driven creator who owes us exactly nothing in explanation, but if a hypothetical creator would prefer we at least watch the damn video that can be an entirely reasonable cost of entry. I guess my question was whether it was intentional! Sorry if my OpenBSD comparison clanged. That OS is righteous.

I don't find the OpenBSD page ugly at all. I find it efficient (and I'm currently not an OpenBSD user or developer).

If you would like to see ugly pages, there's probably a Geocities mirror up somewhere which one can browse through for an ample amount of examples.

A more accurate, descriptive word is, I believe, pre Web 2.0 (or Web 1.0?).

The LibreSSL pages, which are clearly familial, for a long time included "This page scientifically designed to annoy web hipsters". That the design speaks to you says they've got the tone just about right. :-)

That quote refers to the intentional use of Comic Sans, not the overall site design. The OpenBSD project is semi-famous for intentionally using Comic Sans to repel people who seem to be more interested in fonts than in writing code.

We seem to be angrily agreeing.

OpenBSD webpage looked just fine in 1998, and it simply hasn't changed since (why should it?)

Because some of the things that have changed in web page design since 1998 represent actual progress with respect to UI and UX.

They're using modern markup, such as header, nav, and footer, to better represent screen readers. There's more use of color as a secondary identifier, while keeping position and size as primary identifiers. That's lots of improvements!

Well, patience is key. Just read the transcript and you'll notice that the project is not finished.

I think it would be fine if the transcript had photographs, but as it is, the transcript is not really sufficient on its own, so you're forced to watch the video anyway.

or just go check out the comments on HN and see what everyone else says

I think the major difference is that with OpenBSD it's quite clear what it is, an OS so secure that it sacrifices comfort. That's not an option for most people using a Client and probably neither on the Server because those are usually deployed at work places where standardization is usually more important than sophistication. But I agree, it's not clear what NODE is for, although I get the impression it's a remake of 90s credit card sized "PDAs" (not sure if they ever had a real name).

The first two paragraphs seem to summarize the project fairly well. The text itself also appears to be fairly well organized.

I haven't watched the video itself (I'm browsing this between compiles and integration test runs), but I certainly can't find any fault with the textual side of things.

The page is ostensibly a transcript of the video.

I understand in some respects, but for me I just read the headline and go "Oh, that's interesting!" and then read it or watch the video. If I do not like the headline then I do not read the article.

I do agree that subtitles should be included though, basic accessibility in my eyes.

The OpenBSD web page is 126kB (less after gzip). Once the image resources are cached, it's only a measly 4kB for the HTML and 4kB for the CSS. I'd say that more than makes up for the page being ugly.

Looking like it was made by software engineers who don't care about--and perhaps even distain--emphasis on visual design is part of the brand!

I mean if you gave a front-end whippersnapper a 126kB budget clearly something with some design harmony could be made that wouldn't, as I say, scare people off. It sets the tone and in a rather quaint way says "This might not be what you are looking for." I like it.

Now this, this is a project we can all get behind. Adding radio with kill-switch and getting PostmarketOS working on this will essentially make this a smartphone replacement to reclaim mobile computing from the duopoly.

Also, N-O-D-E publishes all the requisite files; Essentially making this a 'Build your own smartphone' project and those who don't want to get their hands dirty could order one from the website when available.

Absolutely nothing will make it a smartphone replacement unless it has good battery life. Smartphones are batteries with computers attached.

> will essentially make this a smartphone replacement to reclaim mobile computing from the duopoly

I think it’s a bit too thick to reclaim anything, to be honest, but its nerd cred is off the charts.

> I think it’s a bit too thick to reclaim anything,

Current chase for thinness by manufacturers, even if it means sacrificing structural integrity[1] is laughably absurd.

May be I'm bit biased since I was frankensteining a similar concept with my limited capacity, but if the thickness of the device in this prototype is ~2.5" to 3" then the size of the final production device wouldn't be a deal breaker for many.

[1]https://youtu.be/eweyhaE6mpI?t=518

I assume so;

I would not replace my proper well working phone with a makeshift device like this at all.

And there are plenty of options out there already; No need to wait for this setup.

>And there are plenty of options out there already; No need to wait for this setup.

Like?

The only current mainstream Linux capable smartphone with physical keyboard I know of is Planet Computer devices[1]; even there, the first party support is for android and so the hardware is closed like any other android device out there.

[1]https://www.www3.planetcom.co.uk/

I'm not sure if i missed it but you did mention physical keyboard;

Without this physical keyboard, you have phones like https://puri.sm/products/librem-5/#overview

Also there are plenty of non google service based android phones and alternatives.

You can also use Linux on most Android Phones as well.

A Pi Zero or a Pi in general is very very far away from what a Android based Phone does and can do. Alone for Security.

What is your thought why you would assume a lot of people could get beind this setup? I do like the project, its nice though but for me probably for different reasons then for you.

'Without physical keyboard' there are about half-dozen alternative mainstream options now[1] for Google/Apple ecosystem, Thankfully.

[1]https://news.ycombinator.com/item?id=24072868

PinePhone has a keyboard attachment based on N900 keyboard (landscape slider) on the way supposedly. It also runs PostmarketOS, Mobian, and other mobile Linux distros natively so we'll see...

I'd rather have a thick phone than a slim chance of not being tracked.

I'll build mine with a metal casing and a GPS for outdoor shenanigans, as sturdy as an AK-47 and as thick as the SCIP

Are you going to build your own cell towers too?

The top-level comment that was being replied to already specified "radio with kill-switch"

That deserves an applaud, but what about Librem? Aren't they way ahead on that path? Why are you not just using that? Drivers too fresh and battery-hungry for you? But it is going to be 10 times worse with any DIY project...

Not only that but im thinking that this could be upgraded, modded, fixed, full control, privacy and last but not least: no planned obsolesce

I agree. If isn't an operational security risk, can you share pictures of your setup?

I too wish to subscribe to this newsletter.

I find my phone way too thin, I can barely use it as-is. Thankfully I have a thick case which quadruples its depth (to about 1cm), which is very usable and still fits in any reasonable pocket.

From the video, it looks like the Zero Terminal V3 is about the same size as my phone, so it would be great, at least for myself.

>I find my phone way too thin, I can barely use it as-is.

I have heard the same reason from many of my contacts, they even use a finger grip on the back of their cases. Phones have crossed reasonable thinness few years back, thanks to a certain manufacturer who started marketing 'Thinness' but have since stopped once their phone started bending in the pants.

This will be a niche product and for that I’m sceptical it’s sustainable on the long run. There’s no way this would even make a dent to Apple’s or Google’s marketshare.

Sure, it’s a nice device for tech-people, but general consumers will most likely not use this.

Not everything needs to be a product.

>There’s no way this would even make a dent to Apple’s or Google’s marketshare.

That's the bane of monopolies or duopolies; no good for anyone except for the 'Investors'.

What kind of brave new world are we living in, where a one-person noncommercial project need be competitive with gigacorp industrial output to be considered valid, let alone judged on its own merits?

This is a superb project, really nice work!

I kept thinking "oh nice, you'd just want to add a---" and you've already done it! Incredible stuff. I can't wait to see how this progresses. The keyboard design is really neat. It reminds me a lot of the old Nokia N900, still the best phone I've ever owned.

Do you have any idea how much battery life you get?

I still miss my old Nokia E7 phone with qwerty keyboard.

This is impressive work, and I would like to have this, or something like this.

A question is whether the amount of clever effort that went into working around constraints of the Pi Zero, would be enough to incorporate the Pi Zero aspects of this assembly into a single board?

Also, maybe not even an exact Pi, but using a RISC-V core and fully open firmware, even if less powerful at this time?

I actually do like the fact that it uses an existing Pi Zero board, even if much creativity is needed to add so much functionality to it in a small space. It reminds me of gopro cameras being taken apart and custom modules added to make it lightweight to use on small drones. I think this is a developing trend, leading to somewhat of an open source supply chain of creative applications for hardware devices. I think we benefit of the economies of scale of the Rpi boards already, just like we do from individual chips, so these custom boards ties that together nicely.

Is there a RISC-V SOC with HDMI?

I am puzzled why there are no widely available laptop-like cases for the Raspberry Pi, especially considering that now there is a 8 GB version of it. "pi‑top [3]" came close, but it was somewhat childish and apparently it's not sold anymore.

Part of the problem is how thick the Pi's are. They're small computers on their own. But in a laptop case, they're chunky. Much thicker than your average off-the-shelf laptop. Plus, there's still rough edges when trying to use them as desktops. The Pi Zero would work in terms of size, but is very underpowered for desktops. The Pi Compute would probably be the perfect fit, but that hasn't been updated in a rather long time. Also underpowered.

So I can see why there's not much of a market. The hobbyists are largely doing something other than standard desktop use, and the people who need laptops (even cheap ones) are better served with something off-the-shelf. The best bet for the future are the handful of laptop-style docks coming out. Just plug your phone or pi into one of the external ports and go.

I would be fine if it was bulkier than modern ultrabooks and more like the old-school Toshiba Libretto. The added volume can be used for batteries, overall sturdiness, as well as other creative uses.

P.S. I decided to try and goodle that topic again, and I found one that is coming soon: https://news.ycombinator.com/edit?id=24092182

Maybe even a card like case that mounts to a case?

I would be ok with a "pi bump" on the case.

6:55 haha one pixel is a letter, the form factor is really cool I thought it would just be a Pi inside but there is also a bigger board combining functionalities.

I do have a inclination for "disposable" computing (old $50 Chromebooks using i3-wm) though and mostly I notice it's the ram so yea i3-wm is a must. This is cool it's full/has everything.

Needs more compute I think. I've started seeing small form factor builds like this 6GB ram 4K capable cube thing for gaming using Atom processor. Oh Larkbox it's called

I'm a bit saddened by the lack of built in keyboard in this iteration. I see the modular bit, but it kinda takes away from the "terminal in your pocket" idea.

The sliding keyboard in the video is only like half a centimeter thick; plenty of room in your pocket with for the unit and attached module.

Now imagine this, just with the Compute Module in a version with proper PCI-E... sigh.

Well, supposedly[0] the next iteration of the CM will be based on the Pi 4 guts, and expose the PCIe capabilities. The maker of the hardware being discussed also mentions that they might try the compute module (it probably makes a lot of sense from the custom PCB side of life?).

0: https://www.tomshardware.com/news/raspberry-pi-nvme-support-...

Oh finally. That's good news especially for people who'd like to tinker with advanced stuff requiring PCI-E chipsets who don't really have the nerve to risk wrecking a hundreds of dollars worth of x86 mainboard/CPU...

> tinker with advanced stuff requiring PCI-E chipsets

What sort of things are you talking about here, out of interest? The only way you're going to wreck hardware is if you're attaching something completely incompatible at the physical level (wrong levels, shorts, etc.); if you're "tinkering" you almost certainly don't have the resources to bring up your own PCIe PHY (there's pretty much no way to do this without custom silicon) so presumably you have some preexisting hardware in mind?

If you just want to play around at the TLP level (for example by using the hard PCIe core in a cheap Xilinx Series-7 chip) then this is all pretty safe: (temporary) catastrophic system instability is one thing (I learned this the hard way when I accidentally had my FPGA spamming MSIs every free cycle), but real hardware damage is pretty tricky.

Gamers and crypto miners do various nasty things to PCIe cards in order to use cheapest cable possible to extend it and actual hardware damage resulting from that seems pretty minimal.

When I said "incompatible at the physical level" I meant it in the OSI sense of an issue at physical protocol level. An electrical specification mismatch could easily cause hardware damage (but would be very hard to achieve unless you were using your own PHY or literally connecting a power supply to the PCIe lines).

> seems pretty minimal

Is there actually any at all? I can't really see how this could lead to any damage.

A Pi4 with 4/8 GB of RAM and NVMe SSD would be a pretty solid general purpose computer honestly!

What kind of SBC to use for such a project really is an interesting question.

On one hand I would love a "shell" where I could just plug in a compute module and have a working portable. On the other hand, the compute modules are more expensive and seem really hard to source, why it's easy and cheap to find a Pi Zero. If one would need to solder it to an expansion board oneself, this would probably be a hurdle for many people. And then, if we are thinking about different boards, why not something else entirely? With more/better capabilities, like included charging circuit or as you mentioned, PCI-E?

I guess there is just no good answer on what would be an adequate basis for such a type of computer. Every option has some drawbacks somewhere.

This seems like a perfect use case for the 3A+ (which is quad-core and only slightly larger than the Zero).

The form factor reminds me of the OQO and similar UMPCs from that era.

Edit: It even has the same sliding keyboard as the OQOs! https://en.wikipedia.org/wiki/OQO

Very nice, I can imagine all sort of applications for this. How much is the power usage?

That keyboard reminds me of the Nokia N900 <3

With that sliding keyboard it reminds me of Nokia N900. I still have one of those .. might dust it off so !!

I would love pocket computer that I could use to programming when I don't have access to laptop.

i miss phones with slide out keyboards

I still have my Android G1/ HTC Dream. If I could get a phone with roughly the same form factor, just a modern CPU/GPU/RAM/camera and bigger battery, I'd be in heaven. Needs to run PostmarketOS or PineOS too.

If he's making PCBs, why not base it on a compute module?

edit: He does indeed say he's contemplating this.

I wonder what the price range for this would be. I’d readily pay up to 300 but would think twice if it goes above that.

im thinking that this device could be upgraded, modded, fixed, have full control over the software, have privacy and last but not least: no planned obsolesce

I wonder if that keyboard module could be reworked to attach to a pinephone; obviously you'd have to replace the actual connection to switch from usb to i2c, but otherwise it looks nearly drop-in.

I'd buy this.

love your videos

Reply on news.ycombinator.com