Slashdot Mirror
Reporter Pans Open Source Laptop Kit TERES-I (theverge.com)
The Verge's Paul Miller has some harsh words for the $242 open source DIY laptop kit TERES-I from Olimex.
Instead of buying one hyper-integrated board that has all of the laptop's brains and I/O on it, you buy several little boards and wire them together. Then you put them inside a mostly finished case built by Olimex -- although if you want to go ultra DIY you can 3D print your own case, too. Everything, from the shell's CAD design to the motherboard's wiring, is available on GitHub for perusal or modification, and the modular nature of the internals means you can add a more powerful chipset or modify just about anything you find unsatisfying about the computer if you have the know-how or if Olimex or others offer compatible parts.
But, unfortunately, almost everything about this laptop is unsatisfying right now. It runs a quad-core ARM64 chip, though x86 and MIPS chips might be offered later on. It has a tiny 11.6-inch screen, a huge bezel, a tiny trackpad, a cramped-looking keyboard, and a whole lot of plastic. The OS (Linux, naturally) runs off a microSD card. At least the LCD comes in a 1080p variant, because the default 1366 x 768 resolution is a real throwback. There's even 802.11n Wi-Fi, which has me questioning what decade it is.
But are there any better alternatives? In the comments share your own thoughts about open source laptop kits.
But, unfortunately, almost everything about this laptop is unsatisfying right now. It runs a quad-core ARM64 chip, though x86 and MIPS chips might be offered later on. It has a tiny 11.6-inch screen, a huge bezel, a tiny trackpad, a cramped-looking keyboard, and a whole lot of plastic. The OS (Linux, naturally) runs off a microSD card. At least the LCD comes in a 1080p variant, because the default 1366 x 768 resolution is a real throwback. There's even 802.11n Wi-Fi, which has me questioning what decade it is.
But are there any better alternatives? In the comments share your own thoughts about open source laptop kits.
I'm struggling a bit with the comment that "1GB is in fact completely unacceptable."
At the risk of i) showing my age and/or ii) getting laughed off the page... I started my career in technology being paid to write software for the 1980s era BBC Micro, a computer that shipped with 32Kb of RAM, of which only 27Kb was usable in the best possible scenarios, and which disappeared rapidly if you wanted anything as high-spec as a graphical display mode...
But behind the ridicule I expect the above comment to attract, I think there lies an important point. Most of us today experience an entire technology stack that has been developed in accordance with some of the rules personified by Eric Raymond in The Art of Unix Programming, specifically things like, "Rule of Economy: Programmer time is expensive; conserve it in preference to machine time". Or "Rule of Generation: Avoid hand-hacking; write programs to write programs when you can"
As a result of this, the technology we use gradually loses sight of the purpose for which it was created. I use a word processor because it is a quick and simple way to allow me to edit a document, layering my thoughts, editing content until I am happy with it, without having to re-type it from scratch each time I want to make a change. There is/was an extremely capable word processing application called Wordwise [which shipped on a ROM chip] for the BBC Microcomputer and which took no RAM [because its code executed in ROM] and which allowed me to edit and maintain documents. Sure, Wordwise doesn't have the features of Microsoft's Word 2016, or LibreOffice Writer 5.0.3.2 [both of which I use], but it gave me word processing with a fraction of the resources demanded today.
I think that we sometimes lose sight of the absolutely insane improvements in system performance over the last 20-30 years - and the complete lack of progress that we see at the human interface. My suspicion - going back to the works of Eric Raymond - are that our developers are writing code that is increasingly inefficient, that the environments that run that code are increasingly wasted [do I really need an animated "ribbon" in my Word Processor - i.e. something that actually slows the software down? No.].
Today we find ourselves arguing that a computer with more than thirty-two thousand times the capacity offered by that fully-functional 1980s BBC micro is "completely unacceptable."
Let's just pause for a moment and consider whether today's 1Gb system is north of 30,000 times faster, better, or cheaper than that 1980s system. Today's machine will surely have many improvements over such early-era systems, but will still fall far short of the orders-of-magnitude improvements that simplistic comparative analysis would suggest. Why is that? Because we have become lazy and inefficient, and so has our technology.
In other words, "If you can't do it in 1Gb of RAM, you are doing it wrong."