Slashdot Mirror
PengPod Crowdfunding a Tablet Made With OS-Switching In Mind
PengPod is running a crowdfunder to create a GNU Linux/Android tablet, the PengPod 1040. This is their second such product; the first was mentioned on Slashdot last year. PengPod has pledged to make all source and tools used to build the images available, so users can build their own OS top to bottom to guarantee that it's free of NSA tracking. The PengPod has previously found some success as a low-cost touch platform for industrial/commercial control systems and is partnered with ViewTouch, the original inventors of the graphical POS to offer PengPod1040s as restaurant register systems. The feature that the developers seem keenest to emphasize is that the PengPod is built to run conventional desktop Linux distros without special hacking required; Android is the default OS, but it's been tested with several others (including Ubuntu Touch) listed on their Indiegogo page.
Saving a spreadsheet downloaded in Android, then rebooting into Linux to edit it, then rebooting back into Android to send it back with a touch-centric email client, or any similar workflow seems rather archaic. It's like over a decade ago when on occasion I would absolutely have to reboot into Windows to get something done - it was annoying (and destroyed my uptime!). Further, you need a keyboard and mouse to operate this device under Linux - at least that's what the demo video shows, and I don't see how it could be done any other way (in a practical sense).
This is the total wrong approach. Now that Android is making it's way onto desktop type machines, perhaps it's time to look at porting applications like LibreOffice and Gimp over to Android. You would still want to plug in a keyboard and mouse to use them, but at least you wouldn't have to reboot multiple times in order to complete what amounts to a single task. Also, it would be nice to drop my phone into a dock that connects it to a monitor, mouse, and keyboard - but only if I have desktop applications available. If for some Android OS-centric technical reason, porting full scale apps makes no sense, then it's time to look at some kind of practical OS hybridization. Not to mention the time is now to steal Microsoft's thunder and if we don't make such moves now, I can't promise it will still be that time in two or three years.
Brought to you by Carl's Junior.
The project doesn't make any mention of the NSA. That's timmy inserting irrelevant editorial.
And isn't even a valid point -- no one can audit every line of code in every piece of software they run - the linux kernel itself has over 5 million lines of code, and that doesn't even include applications.
Even if you were certain that the code itself was clean, how do you know you can trust the firnware in the device, the compiler you're using to compile the code, and even your own computer?
If you went back to a 1960's era wirewrapped computer, you might have some hope at validating the operating system and hardware, but there's no way for an individual to be 100% certain that a modern computer is free of software and hardware back doors when the bad guys could have compromised the hardware locked inside of the core chips that run the device.
VM's have a performance penalty. What's the state of virtualization in the ARM world anyway? Mere emulation similar to what you find in the Android SDK will run faster, but you won't get the full Android experience. Running Linux in a chroot inside the Android host might be a better option performance-wise, but the Android kernel and userland have quirks not found in desktop Linux.
Why run a VM when you can just switch OSs?
Because rebooting any time I want to do any of the following sounds a lot less convenient than hitting a button to switch from the host environment to the VM and back again:
That way if someone emails you a document or presentation you can save it to your device in Android, reboot into Linux and run LibreOffice or another office app to edit it. Or if you’re using Ubuntu and want to watch Netflix or use Skype, you can reboot into Android to do that.
I'm part of the PengPod project, I thought I should come to Timothy's defense. We throw the no NSA thing around in the marketing because we directly share, or use common open source distributions that share, all the source in the machine. For the bootloader, u-boot, kernel up to the user space. We share all the Apache licensed parts as well. In this way, you can review and audit all the code, unlike with many Android devices where their Android modifications are hidden. Of course, as the above points out we cannot actually audit the entire existing code base. Also we don't have source for the 3D acceleration libraries but they aren't required for any uses. Since we provide all the tools to create your own images, the device can be flashed without those.
I own the PengPod 700 and I contributed to this project for the 1040 for an upgrade. Here are some comments I have:
First, the PengPod 700 is great for what I need it for - which is mainly a mobile tablet that boots Linux for taking notes in Vim. I would do the same with a Rapsberry Pi, some USB display, and a battery, but it is all there in one package with the PengPod 700. It fits just fine in a case with a mobile keyboard. Some downsides to the original design: low processor speed + RAM - starting Firefox takes some time, but it still works. I feel like USB in Linux without autologin is a single point of failure - and it has been reported that the connector does suffer. It also doesn't have access to the backlight PWM/GPIO out of the box - so no brightness control, which directly affects battery life, which isn't that great. For what I use it for (taking notes at meetings or on the go), the issues aren't too bad. At $100ish for a Linux tablet, you can't go wrong if you set your expectations right.
Now, I want the 1040 because the specs are amazing for that pricepoint, especially with Linux. I would up my usage of it to playing some light games, spreadsheet, general web browsing - it would really be something that I wouldn't feel bad using from the couch or pulling out at a conference. I still would prefer multiple USB ports, but most tablets don't even have one.
I really wanted an Ubuntu Edge, but didn't really need to replace my current smartphone and honestly I could see myself using the 1040 a lot more. I probably still won't be doing heavy development on it or running WINE (both due to ARM), but I can't really find much that is cheaper from a mobile perspective with the full package running Linux.
I'll respond to this in two parts, first about the old devices. We sold a lot more than 100 in the 6 months or so they were generally available, so it wasn't 1 order and done. We were told we would have some notice when they were going away so we would have a chance to 'stock up' based on our estimated sales volume. That didn't happen and in fact our last regular order was cancelled and refunded after being delayed several weeks. We had some alternate choices but everyone was switching to a new touch screen control and nand chip but no one had working source for them. We didn't see enough demand to justify reverse engineering the drivers. These things happen at small volume and we no longer using the original supplier.
Second, about the 1040. We have to order these in larger lots, so the good news is we will generally have a larger supply for fewer surprises. Also at this volume they are being manufactured, PCB and all in custom runs. So as long as the A31 is available, we can have them made, assuming their is demand. A finally because our new supplier actually creates and shares the source improvements needed to run the device, even if small changes come as peripheral chips reach EOL, we will have their source for the new parts directly from our supplier.