Hacker News new | past | comments | ask | show | jobs | submit login
Crowdfunding the Novena Open Laptop (bunniestudios.com)
292 points by kevs on Apr 2, 2014 | hide | past | web | favorite | 74 comments

Having worked with bunnie and xobs at chumby, one thing I can vouch for is that they intimately understand the risks and challenges of actual real-world hardware production at scale.

The fact that they went ahead with the crowdfunding on this is a really good indication that it will actually ship (if the goals are hit), and with a much higher probability of being on time than your average Kickstarter hardware project (where the people behind it usually have never had hands-on experience with actual production runs).

He's also been kicking the project around for years now, so he's probably thought/worked through a lot of the initial hurdles.

At 2:15 in the video, the screen behind his head flashes a loop of the string "Y3Jvd2RzdXBwbHkuY29tL25vdmVuYS1wdXp6bGU=". When run through a Base64 decoder, it outputs "crowdsupply.com/novena-puzzle". It would appear as though we've got an ARG on our hands.

I'm always interested in open hardware but I'm a bit baffled by the choices that are available:

Just the board: This is perhaps the most appealing option but I would want a case with it, I'm not interested in building my own case. Why would I want a 4GB microSD with that? I'll buy my own with a decent capacity/class or even an SSD.

All-in-One Desktop: $700 more than board for a case and LCD? The design of the flip-top LCD doesn't make a lot of sense - you lose a lot of desk space and the board is going to collect dust. It's not very portable due to the lack of a battery, controller board, keyboard and mouse so I might as well just plug it into my monitor.

Laptop: $800 more than the desktop option for a battery, controller board and SSD. Seems a little expensive and again still not very portable. I'm not sure you can call it a laptop without a keyboard or pointing device.

Heirloom Laptop: $3000 more than the "Laptop" for what is now actually a laptop that includes a keyboard and pointing stick with a wooden case. Wood is quite heavy so it's not exactly the ideal material for a portable device.

The first thing you’ll notice about the design is that the screen opens “the wrong way”. This feature allows the computer to be usable as a wall-hanging unit when the screen is closed. It also solves a major problem I had with the original clamshell prototype – it was a real pain to access the hardware for hacking, as it’s blocked by the keyboard mounting plate.

The point of the thing is to be easily modified, and the design facilitates that. If you don't want regular access to the components, you can leave it closed and treat it as more of a conventional all-in-one.

You could conceivably install the LCD the other way, too.

Sounds like you just want a regular laptop, there are many companies who already make those.

My guess is that he's sourced an incredibly cheap batch of 4GB cards from China. It likely adds very little in terms of cost, but gives much more value to the end-user in being able to just boot it up and have it work.

You are correct, the 4GB SD card is mostly about having a quick start ready-to-roll image.

Bootstrapping SoCs gets really tricky when you don't have a JTAG device. The newer chips that let you boot directly from SD have been a great leap forward in development speed.

So wish I could upvote you multiple times. People these days are spoiled with their precompiled ROMs and precompiled software that runs on mainstream desktops to load those ROMs over universal standards (USB). Even with JTAG embedded programming was a PITA.

You want to talk about the days of erasing EPROMS in a UV oven and waiting 15 minutes for a 27C512 to burn? =) Because that sucked.

If you wanted just the board + case, and aren't really interested in hacking on the device, you might want to look into the Cubox-i4pro which uses the same cpu (though slightly slower clock, 1000 vs 1200 iirc)


I just got my pre order a few weeks back, but haven't had the chance to play with it yet.

It looks like the Cubox isn't open. At least, it fails the 10 minute rule. I couldn't find the design files, beyond pdfs for the schematics of the "upper" and "lower" PCBs (which don't meet the criteria of being open) and nothing for the main board.

It really isn't open hardware... but it's pretty much the same cpu/chipset and if you're fiddling with software, it's a much better option price wise. That said, TFA's open laptop really isn't so open either, considering the gpu/accelleration abilities are locked up. You give me open hardware with open drivers at something resembling cost competative, then I'll take notice.

Thank you, bunnie, Xobs, et al., for listening to the interest that was expressed for what was a personal project and deciding in response to expand its scope (quite significantly), providing this momentum to open-source hardware and systems.

(I don't know bunnie nor Xobs, but this is what I gather happened based upon reading occasionally about the project over the past some months.)

From my perspective, bunnie provided a write-up of what he was up to, purely for the interest of those who cared to read. There was overwhelming response.

If you care about open systems and hackability (in the classic sense), this is really worth looking into and hopefully supporting. (From my admittedly limited perspective.)

Like the Open Pandora (I still use every second day) or Openmoko (paperweight), I'll happily put my money forward supporting open hardware.

Congratulations Bunnie/xob/others on the release of the crowdfund.

This is stage one. Community, hackability and continuous development cycles will lead to better, more feature packed and stable devices. The CPU was a good choice for performance/power/openness but there is nothing stopping them from using a modular design for the SOC and people could replace them as required.. the www.pyra-handheld.com/ is build on that idea.

I'm looking forward to the use cases for the laptop...

Side question: Is there much in the way of design that can be done to improve the physical security to this design? or are we hoping that people simply don't have the driveby ability to mess with this due to the fringe-ness? and physical access is physical access.

For those that didn't instantly recognize the name, this is Bunnie Huang's open laptop he talked about previously [1]

[1] http://www.bunniestudios.com/blog/?tag=novena

I have been waiting for some time for the crowdfunding campaign.

The FreeScale i.MX6 platform is very versatile. Up to quad core 1.2GHz, SATA, PCIe, Gigabit Ethernet (limited to 480Mbps) and lots of good documentation and driver support.

The only issue is the graphics blob and... in this case the price tag. I hoped I could afford it, but it looks like I can't.

Still, there are lots of other community boards based on the i.MX6 SoC. If you don't need the laptop version or an FPGA, there are much cheaper boards out there.

The Novena Open Laptop really is good work. I hope it'll get funded!

Open laptop because we're worried about state-sponsored attackers expending multiple millions of dollars of time and resources to undertake the lowest level of attacks upon our hardware.

Top model ships with a keyboard with a radio transceiver built in.

Defending against an APT is not the only use case for the machine. More importantly you singled out the top of the line model for your complaint. For that much money it better come with keyboard. Worried about APT? Save thousands and build your own ergodox to use with the novena.

I'm happy to see some sane crowdfunding pricing on these reward tiers. So many projects have their reward pricing set similar to retail prices. That's fine if you've already done all the R&D work on the project and there's little risk, but if you actually need net funding from the project then you need a significant overhead on every reward.

That said, I hope this is successful, these guys are consummate hardware hackers, it'll be neat to see what happens when they ship a project on this scale that ends up in a lot of people's hands.

They sure made it sound liked the crowdfunded prices were cheaper than retail: "We are offering four variations, and at the conclusion of the Crowd Supply campaign on May 18, all the prices listed below will go up by 10%"

I read that as the post-campaign follow-on pre-orders which tend to be fairly standard these days.

The boards are a little out of my price range at the moment. But definitely wanted to support this project, so I have backed "Buy us a Beer" at $5.

Love at first sight. The only thing left is deciding whether to buy the desktop or the laptop version.

This is GREAT to see Novena make it to this step. A truly awesome amount of work by bunnie and xobs it exciting see them make this far!

this guy is pretty awesome, might be just the hack candy one might like to have if one is into hardware hacking.

TL;DR: here is the video from a conference where bunnie and his friend hacked flash card controller, so you can see the qualities that the laptop will possess.


Why use (and advertise) "High Modulus Epoxy" as the bonding agent for the wood version? High Modulus means stiff, and unbending - i.e., crack instead of bend, and in bonding wood to aluminum, a lower modulus resin would be better for this case, as it would flex when stressed and not cause a snap-off delamination?

Two possibilities:

1. The guy who designed the wood version is experienced in bonding wood to aluminium and has good reason for using high modulus epoxy.

2. "High Modulus" sounds really good, let's say that! Nobody wants "low modulus epoxy", whatever that is!

What are the use cases for the FPGA realistically? Then only piece of easily downloaded and installed code is a SHA based crytocoin miner.

Actually pretty handy for doing optimized NIC/protocol processing. Particularly handy for wireless stuff IIRC. It's about as close as you can get to a software defined CPU, so the options are kind of limitless.

Learning & prototyping hardware (ASIC) design. Applications that like low latency I/O. You can respond in microseconds with real time guarantees, eg for radio link layer protocols.

Just to toss out a few ideas: software defined radio, high-throughput DSP (like audio processing or machine vision), MITMing of fast protocols, packet inspection, CPU emulation.

I found http://opencores.org/projects nothing is a killer app right now.

At work I use an FPGA as async real time controller. Not sure what utility that would have if not connected to specialized hardware.

A little off-topic, but they gave a nice talk about SD Card exploitation at the 30C3 in germany: http://www.youtube.com/watch?v=CPEzLNh5YIo

Not really off-topic: at 42:00 the setup shows a Novena prototype. They used a binder as case.

Great idea really I wish them success!! but why are they using the keyboard from a Thinkpad.They are the most infuriating keyboards ever as the fn key is on the bottom left side with Ctrl one key in one. Even Lenovo's come with a switch in the BIOS to undo this. Before anyone says this is the actual format of the qwerty keyboard the overwhelming majority of people are used to it with Ctrl on the bottom left so just leave it

I'm not so sure the majority of developers / hackers (the target market for this device) are used to Ctrl on the bottom left. I know several (myself included) who remap their keyboard so it sits left of "A".

Thinkpad keyboard only appears to come with $5,000 limited-run version, and it's detachable. Every other version explicitly does not come with a keyboard because of how little consensus there is.

Very curious about the capabilities of that FPGA in combination with the rest of the hardware, and the high-speed expansion port.

Oh yeah. I've had my eye on this for ages. It comes with an FPGA!

If Matt says it's good, then it's good.

Somebody please make a device which is more accessible (pricewise), and can be used to prototype/interface with common digital protocols (i2c, uart's, spi, CAN, ethernet, USB).

I'd imagine it could be a mashup of SBC (Beagle, Olimex's LIME, heck, even RaspberryPi) and Bus Pirate. Add a screen and battery, and call it a day.

I don't think a device like this should have the fastest CPU and plenty of RAM (that wouldn't hurt of course).

i'd really like to get the main board _and_ the battery board, but no display (wrong size) and no case (wrong size). i'd like to use a modded thinkpad x-series case (and keyboard) and install a better display and board!

Some interesting choices, some incredibly disappointing.

A9 core rather than the virtualization-ready A15? (I'd rather have a 64-bit ARM but those are still a ways out)

Edit: and let's not hear about how closed A15 is blah blah. At the end of the day, something is closed. Did Freescale give you silicon masks? Guess it's not totally open after all.

Single SO-DIMM means max 8GB of RAM, but of course since you're running a 32-bit chip you'll be restricted past 4GB anyway.

100 Mbit Ethernet in the Year Of Our Lord Two Thousand Fourteen? Edit: Several people have pointed out that, thankfully, the 100Mbit port is a secondary port, while the main one is Gbit. This is actually pretty cool.

The case is actually kind of cool, but you can't really call it a laptop if your footprint is normal laptop size + a separate keyboard. Nobody's lap is that big. I think it could be pretty usable for work while on travel but you can't use it on an airplane tray table.

And then the pricing. Oh my the pricing. $700 for a case and a screen. Another $700 for an RC car battery and an SSD. Not even Apple marks up SSDs and batteries that much.

I would really love to get the laptop spec one for $800-1000--you know, the price of a great desktop PC plus a monitor. Sure, the ARM processor is going to be disappointingly sluggish and I'll be beating my head against the 4GB memory ceiling within an hour, but it looks really damn cool and the screen size is acceptable.

Kickstarter-type things are often done as "backers pay more for a unit, but they get it first and they're willing to pay because they believe in the product". However, the prices will go UP by 10% after this campaign is over! So the $2000 "laptop" will now cost $2200, despite the reduction in component cost due to time and scaling.

Edit: oh, and apparently the case, which I think is one of the cooler parts of this project, doesn't seem to be open? So I can't pay a fabricator $200 to make one, then slap a $400 A15 board in there. The totally open laptop, ladies and gentlemen!

I know Bunnie as an incredible hardware engineer with a lot of experience across the entire process of getting a product from design to manufacture. I'm sure the decisions were made for very good reason.

Freescale has extensive open source board support packages for several flavors of Linux and Android. With the exception of the TI's OMAP5, all of the A15 implementations I know of are under lock and NDA and even though Bunnie could source them, the documentation and software support packages would not be as easy and would require working closely with the manufacture to get them. I have no idea about OMAP5 as I haven't even seen one in production. Most mobile ARM processors AFAIK do not support gigabit ethernet yet, save a few Marvell SoCs (less for mobile and more for other embedded).

Smartphones designed with modern ARM processors require teams of 2 dozen engineers working their asses off for months to make a single revision, and this is when they're partners with the chip manufacturers (ala Google, Apple, or Samsung). Bunnie is working with a much smaller team and with different constraints.

Any board like this is going to be expensive to manufacture and assemble. Even if he were to do a run of 100 units at a time, the overhead of assembly and part purchases in such low quantities are likely to run $100+ a unit. Combine with the extra support costs of including an FPGA in there, and it's not unreasonable to charge so much for such a niche, low quantity product.

Edit: Also Freescale is on the low end of the high end ARM market and they know it. I wouldn't be surprise if Bunnie got some extra help from them during integration.

I wouldn't be surprise if Bunnie got some extra help from them during integration.

I've always been happy with the amount of support Freescale puts behind their processors. You don't realize how bad it is with some manufacturers until you're halfway through a project and things just fall apart.

And the i.MX6Q is a snappy little processor for the price.

I think the OpenPandora project has confirmed the OMAP5's availability and has samples.



Except for TI is kinda 'checked out' of the Linux support business sadly.

Are they expecting their SoCs to run Windows instead?

No. They seem to be focusing more on the RTOSes, like VxWorks and QNX. There's more to OSes than just Linux and Windows.

What Sanddancer said, they are investing in their Tiva (aka Cortex M series) parts and de-emphasizing the OMAP line. (they may have said the OMAP5 is the last of that series) I don't think they felt they were offering a lot of value over the other Cortex A9 vendors.

All of the enclosure design files will be published on the Wiki shortly, so you'll be able to make your own. The desktop/laptop versions ship with a spare bezel, as well, so you have one to modify. You could happily make your own case or pay a fabricator to do so, but the case in the campaign is injection molded, so you'd likely pay a lot more to do so.

The focus of this laptop is really on versatile and hardware-hacking-power, not on raw performance. There are a number of flagship features that you won't find in a typical laptop, and gobs of expandability.

For example, there's an onboard FPGA, for when you're at the coffee shop and you need to MITM a high-speed / low-latency protocol.

The schematics are full of attention to detail and fantastic surprises: http://www.kosagi.com/w/index.php?title=Novena_PVT_Design_So...

As you've pointed out, it seems like a big price jump to go from the bare board to the desktop and from the desktop to the laptop, but you're actually getting a lot of components there, including: very flexible battery charge controller (runs ChibiOS on an ARM7), speakers, speaker mounts, machined SSD mounting rails, and cabling. The screen is a gorgeous IPS panel.

Also, regarding the airplane tray table: it's actually designed to be hung off the seat in front of you, so you've got the entire table free.

> A9 core rather than the virtualization-ready A15? (I'd rather have a 64-bit ARM but those are still a ways out)

There aren't really many choices. Samsung has A15 cores, but it's probably hard to get and there's _no_ documentation. The same for Nvidia Tegra. Not sure how easy it's to get OMAP5, and how expensive.

The i.MX6 SoC is probably the developer friendliest one, has very good documentation, and it has support for lots of stuff.

> 100 Mbit Ethernet in the Year Of Our Lord Two Thousand Fourteen?

Read carefully. One of those ports is 1Gbit (due to SoC limitations only 480Mbps though). The other one seems to be connected via USB 2.0. One could have put an Intel PCIe Gigabit Ethernet chip on the board, but the PCIe slot is probably better used for WiFi.

It's pricy, but I don't think they are able to produce them for fewer money, like HP/Samsung with their $400 Chromebooks.

This is not a commodity, value product. The cost is reasonable for the short production run. The FPGA is a unique feature, and there are plenty of other points you can extend from. If you're not building your own hardware and integrating it with the laptop then it is not for you and you would be wasting $2000.

(Also: "no wireless, less space than a nomad. lame.")

The 100mbit ethernet is a secondary port. It has two, so you can act as a MITM or router or whatever. (search the crowdfunding page for "1gbit")

Thanks for pointing that out, that's a relief!

The price is an issue here.

I understand it's supposed to be open for hardware hackers, but lets face it: who wants to hack a laptop? most hardware hackers are happy using a raspberry pi and a Ubuntu laptop. Why would you want to risk breaking a $1k open laptop to open your garage door when you can get a fully functional Linux computer for $35, or a cheap Android for $60?

who wants to hack a laptop?

I do, kinda sorta. I do 99% of my hacking on a Lenovo W500. I have some desktops machines but they serve more as media devices and for background processing of a few things.

Back when, I used a desktop for almost everything. The nice part was I got to build a machine that suited me. If something new and better came along (CPU, graphics card, monitor) I could upgrade my system piecemeal. I typically didn't have to give up the screen I like (unless I was swapping it for something better) or lose the keyboard I've become familiar with.

But now, using a laptop, I'm stuck. My Thinkpad is getting a bit dated. I'd like to upgrade, but buying a new laptop is mostly an all-or-noting affair. Want the higher rez on the new Thinkpad? Well, then you have to take the goofy giant trackpad and lose the dedicated trackpoint buttons. And there's a numpad you'll never use, too, making the main keyboard off-sided. Want a different keyboard layout? Well then you can't get the nicer rez or 16GB of RAM.

In an ideal world I'd be able to upgrade a laptop the way I've upgraded desktops: Better CPU available? Swap it in!. Nicer keyboard? Go use it.

Ideally the display, main CPU/motherboard, and keyboard would all be separate things. I long for the day when my "phone" is powerful and flexible enough that it can be used as a "laptop" by way of wireless keyboard and larger display.

The laptop under discussion isn't for me, but I love the idea of it.

"Novena is a 1.2GHz, Freescale quad-core ARM architecture computer"

That's a disappointing choice! Why not something more like LEON or OpenRISC?

How about finishing some projects before starting new ones?

I'm thinking of e.g. NeTV

http://www.adafruit.com/products/609 http://kosagi.com/w/index.php?title=Main_Page

What's unfinished about the NeTV? You can buy the kit and download the source code and do want you want with it.

It is totally understandable -- and totally disappointing -- that this custom work of art in hardware comes with regular, bloated, unhackable linux. One can dream..

I feel like you must have (accidentally?) withheld context. Since everything is open, you can presumably produce drivers for QNX, FreeBSD, or another OS of your choice. I'm sure you can also boot bare-metal into embedded firmware of your own design.

Linux can not be completely understood by one man or even a team of them, and by the time you do figure out what you need to edit, the whole thing will have churned.

Linux is a complex operating system, and it takes a genius to understand the complexity.

Yeah, sorry, crappy writing skills. I tried to elaborate in a sibling comment: https://news.ycombinator.com/item?id=7518682.

I'm gradually growing aware that just having the sources be open is not enough to be hackable.

"unhackable linux"? What?

I suppose I should elaborate on where I'm coming from.

1. http://queue.acm.org/detail.cfm?id=2349257

2. https://plus.google.com/110440139189906861022/posts/EKhGyqn1... (including comments)

Don't get me wrong, I use linux everyday and it's better than the alternatives out there. But seeing a blue-sky bespoke attempt like this gets me wanting more.

The first paper is (a) complaining about FreeBSD and (b) he hasn't seen how bad most commercial software is on the source side.

The second is about suspend; while I somewhat agree on not liking high integration (dbus, high-complexity desktops), suspend has been a total nightmare largely due to hardware manufacturers. You need to do very specific things or the display doesn't come back cleanly.

The first paper is not complaining about FreeBSD, it's complaining about open source software in general, using the FreeBSD ports tree as an example. All of the software examples given are equally valid for Linux -- libtool and all its ugliness is still required, a lot of intermediary libraries are thrown out in the end result -- and the lessons are still unlearned. The mess of OSS begat ALSA, obstinate kernel developers refusing to do the right thing with ALSA begat PulseAudio, etc. Yes, commercial software sucks, but that does not excuse just how much of an unholy mess open source software is.

Software architecture is harder than it looks. Where's your alternative tidier solution?

I agree that libtool is ugly, but it lets you build on different systems with different management that are immune to harmonisation. The sound mess is similar: people refusing to give up their own ideas on what constitutes a good solution. We don't have a magic wand to make people reach consensus. The best we've managed is Jobsian totalitarianism, and even that doesn't always produce the right answer.

I'm working on a very radical but very speculative answer: replace compatibility guarantees with tests, try to be harmonious by making it much easier to make changes to code one didn't write. More info: http://akkartik.name/about. But it's likely to be wrong, and even if it isn't, it'll take the rest of my life to convince myself and maybe others.

Specifically what OS did you have in mind?

Registration is open for Startup School 2019. Classes start July 22nd.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact