The sweet Raspberry taste of success masks a missed opportunity
- Reference: 1734339552
- News link: https://www.theregister.co.uk/2024/12/16/opinion_column_future_raspberry_pi/
- Source link:
The full spectrum of geekery, from precocious prepubescents to CEOs of specialist technical companies, has grown to expect an endless stream of well-engineered hardware backed by a mainstream software stack and a thriving ecosystem. From the [1]$4 Pi Pico microcontroller to the new [2]$90 Pi 500 complete desktop computer , Raspberry is what an Apple run by Steve Woz instead of Jobs might have looked like.
As a result, we Pi fans have been spoiled rotten. We don't see the profound challenge of masterning cheap, powerful and manufacturable at scale - the impossible triangle that sank most of the UK's first generation of computer makers. We complain when consumers can't get boards during [3]supply chain meltdown because of the insane success of the Pi in commercial use.
[4]
Success here not only keeps an industry sector alive, it guides the continued relevance of Pi evolution in ways that benefit us all. We look at the Pi 500 and don't see a ridiculously functional, very low budget educational and home computer built around open source, but something unexpandable with missing options and internal tinkering discouraged. Try finding anything near as good. How dare we?
[5]
[6]
These may be the misperceptions of an ungrateful, entitled nerdery too used to a good thing, but there's a deeper truth at work. In part, this is an unavoidable consequence of the Raspberry Pi idea growing from a single, focused product of "a [7]BBC Micro for today" to a revival of proficiency in low-level software and hardware skills amongst a very diverse enterprise that serves many different audiences. In part, this is because, despite the ethos of early microcomputing still fuelling the Raspberry Pi-maker's thinking, it isn't possible to recapitulate the raw excitement that happened back then.
Pi-meister Eben Upton has never hidden being a huge fan of the Amiga 500, which parachuted sound synthesis, multimedia graphics and multitasking into markets that were used to machines which spent most of their time pretending to be glass teletypes with crude bitmap graphics stapled on. The chance to make a Pi 500 in tribute was, he hinted, irresistible. It could never be that tribute, though; the Pi 400 was the bold move into a market that nobody knew could be reanimated by Pi magic. The Pi 500's existence is a strategic success where strategy is the hardest trick to pull off. It is deserving of more attention offworld of Planet Pi than it will get. But what it is not, is exciting. Somehow, we expected excitement.
[8]
In a way, unexciting is good. The Raspberry Pi has already outlived [9]Commodore Amiga , which lasted less than nine years before being sold off, crushed under the wheels of Wintel. Raspberry Pi is its own Wintel; it may not define single-board computing but you need a top-notch reason to go elsewhere.
That first generation of mass market microcomputers created a generation that is confident in its own expertise. It knew what computing meant from the ground up. The Raspberry Pi concept has been beyond excellent at opening up hardware for anyone to do anything with. Moreover, it can do so with the most modern of software - that hardware is coupled to a [10]full Linux environment .
That is essential to its relevance, utility and long life. What it is no better at than anything else, is enabling people to iteratively and effectively learn how to be a modern programmer. In their day, the first generation did that. That got lost in the rainforests of complexity overgrowing in ever more fertile silicon soil. Planet Pi retains all this daunting complexity, even if it does encourage us to descend from the canopy and poke around among the roots.
[11]
Going back to basics in a modern way was always part of the original brief - the Pi was a panning nod to Python, which was to take the place of BASIC as the switch-on-and-go instant gratifier. That concept never really came off. What is needed is the system software equivalent of Lego – where the grimy details of interfacing software components are as hidden as the screen memory mapping arithmetic underneath PLOT X,Y in BASIC.
It is hard to imagine the nature of abstraction that's needed to untangle even a little, limited subset of the structural complexity that modern systems present, and it is absolutely no criticism of Planet Pi that it doesn't exist there. It is the ideal place to start, though, as known to home lab builders who want multiple truly independent machines to explore ideas with. Want a real eight machine cluster to practice your Kubernetes? Well then.
[12]Raspberry Pi 500 and monitor arrive in time for Christmas
[13]Raspberry Pi Compute Module 5 cranks up the power – and the heat
[14]Framework laptops get modular makeover with RISC-V main board
[15]Ambitious overclocker cools Raspberry Pi 5 with liquid nitrogen
Encouraging naive users to go under the hood is always going to be a bad idea on systems with other jobs to do. Nobody wants to build Lego models on top of a running car engine, let alone ones that interface with the fuel injection, explosions and HT leads, even if that's the best way to learn about how internal combustion actually works.
It will take a lot of very smart people doing things for the first time to Lego-ize modern system ideas. That's OK. It took a lot of very smart people to make the first generation of computers work, and there are a lot of very smart people as a result dedicated to the Raspberry. Time to start making the hard stuff as easy as Pi. ®
Get our [16]Tech Resources
[1] https://www.theregister.com/2024/08/08/pi_pico_2_risc_v/
[2] https://www.theregister.com/2024/12/09/raspberry_pi_500_monitor/
[3] https://www.theregister.com/2022/12/12/100k_raspberry_pis/
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/personaltech&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2Z2AIVjK4FuHbq-6fef61igAAAMA&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/personaltech&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Z2AIVjK4FuHbq-6fef61igAAAMA&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[6] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/personaltech&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Z2AIVjK4FuHbq-6fef61igAAAMA&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[7] https://www.theregister.com/2024/01/18/bbc_micro_bot/
[8] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/personaltech&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Z2AIVjK4FuHbq-6fef61igAAAMA&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[9] https://www.theregister.com/2012/01/02/commodore_64_30_birthday/
[10] https://www.theregister.com/2024/03/07/alt_pi_5_linux_distros/
[11] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/personaltech&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Z2AIVjK4FuHbq-6fef61igAAAMA&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[12] https://www.theregister.com/2024/12/09/raspberry_pi_500_monitor/
[13] https://www.theregister.com/2024/11/27/raspberry_pi_compute_module_5/
[14] https://www.theregister.com/2024/11/18/riscv_framework_main_board/
[15] https://www.theregister.com/2024/11/13/raspberry_pi_4ghz/
[16] https://whitepapers.theregister.com/
Re: On a slight tan(x)
I love the Arduinos and the PIs for different reasons.
PIs can run a full blown OS, whereas the Arduinos are very task dependant controllers.. My favorite usage for the Pi is Home Assistant, and I currently have a home made Arduino dongle on my PC for a personal hobby project, which it does just beautifully/cheaply ( it simulates a Keyboard for various reasons).
The PIs are expensive for small projects and this is where Arduinos are perfect. With Arduino one must be very careful with keeping coding to a minimum and not importing just any old library as memory usage quickly fills up.
The Arduinos have limited memory and no user interface ( outside of smaller LED screens) and this is where the PIs are perfect.
I am happy to use both, they are both extremely useful for interesting home projects. .
Not even the A500 was "switch on and go"
You had to boot up Workbench first then find the Extras disk and then find the BASIC icon which meant many people didn't find it. Also there were different incompatible versions of BASIC depending on which version of the operating system you had - for Workbench 1.0 (the A1000 only) there was MetaComCo's ABasiC, for Workbench 1.1-1.3 there was Microsoft's Amiga Basic, then for Workbench 2.0+ there was ARexx so if you wanted to program in BASIC you had to be motivated enough to buy AMOS or Blitz BASIC.
There's nothing like turning on the computer and having to use BASIC to load everything as in the 8-bits. Imagine if you turned on the Pi and the first thing you saw was the Python REPL which you had to use to load the GUI or a program, that would be the only true switch-on-and-go Pi, but it would also shut out a whole load of other people.
Also could we just stop pretending the wedge Pis are named after the Atari 400 or the Amiga A500, that's just marketing. If they insist it's based on a computer instead of the base Pi model * 100 then they're going to have to find a way to make everyone nostalgic about the Amiga A600 which I'm afraid is just impossible.
BBC Micro..
The Pi as a "BBC Micro for today" is one of those weird myths that is perpetuated despite all evidence to the contrary. The Pi was designed first and foremost as low cost accessible hardware, with the software environment left largely to the user and a challenge in most educational environments. In early days, you were on your own until ubiquity encouraged a slew of "get started with the Pi" kits that provided sd cards, cables, keyboards and cases to make sense of that bare board.
In contrast the BBC was expensive and prioritised quality and a complete environment over convenience. From the start it was possible (and encouraged) to put together a standard configuration that gave you an entirely predictable software and hardware environment on which you could learn. Equally, you might argue that the BBC was designed to make the hardware *more* accessible than the Pi - busses and interfaces proliferated, allowing the base hardware to be hacked, altered and upgraded. In contrast the Pi gives you a (very well) curated set of standard external interfaces around a tightly integrated and largely immutable core.
None of those differences are bad - they're a natural consequence of the evolution of the computing ecosystem - but they also reflect different goals for the projects. I'd argue the Pi is no more educational than any other SBC (and there are a lot), and the Pi Foundation have not done as much as they might have to address that use case. Indeed, it wasn't until the 400 that they addressed the challenges facing many classrooms of having a simple 'plug and go' option that helped with setup and cable management.
The PI is incredibly useful, and a huge achievement in establishing a ecosystem that many educators, developers and experimenters have grown to rely on, but it's achieved that through ubiquity rather than some sort of embodied 'educational design'. In that respect, the Arduino project is arguably better at presenting a user with a environment in which they can learn about both hardware and programming with a low cost device.
Re: BBC Micro..
ubiquity was the point - for education it's pretty much the only requirement
The most important design aim was that you could break the OS install and be able to start again easy. In contrast to the complex support required of a desktop Windows setup. Allowing kids to play with and break things, rather than being sandboxed and threatened has been the goal and I think it achieved that.
Made with love
What's rare about Pi is passion for doing more with less. If Apple bought them, it would be over.
I wish the whole economy worked like this. The World would be a much better place.
My son has a couple of Raspberry PIs he uses for various things. My ancient ThinkCentre desktops run faster, cost less, do more, include storage, remember the time and don;t require fucking about with SD cards to get something that will boot on them. The Pi seems to be a nice embedded controller for various devices, like EV charging points, but as an educational tool it's a failure and a waste of space.
"Time to start making the hard stuff as easy as Pi"
And you can start with the incoherent mess that is Linux
Or that sack of shit Windoze.
Cheap
Guess it depends on what and how you're counting, but compared to where they started I don't currently really factor a reasonably capable Pi + all the other bits I require to get it working as 'cheap'.
Repeatable utility embeddable computing maybe, but the days of 'cheap' really seem to have gone.
PS. I don't really factor in the Pico when thinking 'Pi' as it's a completely different type of product.
On a slight tan(x)
Can I give a similar shout out / slap on the back, to the Arduino.
Given a slightly complicated set of outputs to generate every time some input happened, a board that cost the price of a Happy Meal / pint, let me do it in the time I would once have spent working out which precise version of an 8051 I had for the settings in Kiel compiler