Voice of the Masses: Is it time for a new Raspberry Pi?
Without doubt, the Raspberry Pi has been a whopping success and introduced countless people to Linux. But the Model B hardware is looking a bit long in the tooth now, and faster alternatives with more RAM are starting to appear.
Eben Upton of the Raspberry Pi Foundation has said that he won’t consider Model C for at least a year, because he doesn’t want to “orphan 2.5 million people who are committed to the current platform”. Fair enough – he’s trying to avoid fragmentation. But is it a good idea to wait so long? Will the current Pi be more than good enough for the next 12 months, or will the competition start to nip at its heels? And what, in your opinion, should be in the Model C when it arrives?
Let us know your thoughts, and we’ll read them out in our next podcast, due to be recorded in the lovely walled city of EboracumYork.
Related Posts
About The Author
Mike Saunders
Mike stores his data by printing out hex dumps and laminating the sheets. He's the author of The Haynes Linux Manual and the creator of MikeOS, an operating system hand-built from assembly language.
The success of the x86 PC when there was a large array of (sometimes superior) computing platforms, was the creation of standards that other were able to adopt and enhance without breaking compatibility. The rPi has such a potential to create a platform with published standards, that others can enhance with a commitment towards preventing fragmentation.
Sure. I’d like a more powerful RPi. I’ve been playing with the idea of using it as a cheap small-business Asterisk server, but the performance and line quality is wobbly once you get to two active calls, so I’m stuck flogging old NEC Powermate Pentium 4’s instead.
Ideally, a new model would come with PoE and revolutionize the consumer-router market into making it a standard feature. That way I could literally just hook it up to a router and leave it forever. But that probably wont happen, so optical audio out would suffice.
I hope not, I’ve just ordered my third one and could do without any more for now.
There already are more powerful devices. Very few people use them, they are more expensive, and have yet to get a foothold. Power is not necessary to sell a product. Marketing starts it off, public inertia sustains it, and an established ecosystem makes it persistent.
A more powerful device? Sure. Make it basically compatible, stick a more recent ARM chip in it and add PoE.
Focus on “Anything you can do with your Pi B you can do with your Pi C. Just download a different version of Raspberian”.
Clues exist already
Raspberry pi …cost 100% – sales 2 million
Beaglebone …cost 30% more – sales 100,000
Banana pi R…cost est 50% more – recent no figures
Via APC board …cost est 50% more- sales unpublished..minimal
Mele A1000 …cost 100% more – sales unpublished…minimal
Many, many more exist; all will fail unless they beat it on price. The rPi is not meant to be a super-PC. This is LOW COST general purpose device that can used in a wide number of applications, from switching on an LED, to being a hotspot, to being the heart of a robot and yes to do some PC-like activities. Computational power is not the limiting factor for most of these applications. Therefore the rPi will be replaced when it can be done so without altering the price.
This post says it all. The secret behind the Pi’s success is its price. The model C when it arrives should have specs that are an improvement on the model B, but should cost the same as the current model B. Until this can be achieved I don’t see the point of releasing a new version.
I’d pay £5 more for a physically larger device with a less compromised board layout: Sockets that line up and don’t occupy all four sides, no protruding SD card, the P5 holes moved to a sensible location and populated with a header.
With a physically larger board they could also bring out the I2C bus lines to several identical headers or solder pads to allow small boards to be attached without blocking the GPIO port.
Bonus points for adding an audio input or two and a small reset button.
And there will be those who say “I would pay £5 more for board that is more for a board the size of a dongle” (the origical pi was this size) with fewer ports, etc. Everything is a compromise, my friend, and finding the sweet spot that pleases everybody’s diverse needs at a reasonable cost is the difference between commercial success and failure.
Absolutely everything is a compromise, but in the case of the Pi’s board layout I think it was a compromise too far, with little benefit other than being able to say it’s “credit card sized”.
There’s no denying that the board layout has issues – the first release didn’t even have mounting holes! I’d just like to see a model with those issues resolved, which otherwise maintains compatibility with the model B, rather than any major change to the architecture.
To which I’d add an SD card connector not made of cheese. I replaced the connectors on my Pi’s because I was sick of corrupted filesystems. Talk about spoiling the ship for a ha’porth of tar: connectors which work aren’t exactly expensive.
Power corrupts, and computational power exponentially so. It is great to have high spec devices, but it makes you lazy, code inefficiently and less able to cope when faced with situations when you not have those specifications. Working with low specs makes you think more, create ingenious solutions to problems, and target perfomance with clever coding rather than brute force. Come back assembly language!!
No. Power has never been what the Pi was about. It can still fulfull its intended function quite happily and avoiding fragmentation is essential to getting good eductional resources.
As several people have mentioned there are more powerful boards for people with greater needs, but these people are also more likely to be technically adept.
Ebden is completely correct.
Once wayland and maynard are standard the current raspberrypi will seem a lot faster for desktop use. Though rendering javascript heavy websites will probably always be slow.
Taking the ethernet port off the USB bus, making it gigabit and improving bandwidth to USB and SD could make a big speed difference without effecting compatibility.
However a dual-core armv7, with 1 or 2 GB of RAM would mean the raspberrypi could replace a desktop for quite a few computer users.
The new Banana pi (http://www.BananaPi.org) at $49 offers a more powerful dual core processor, 1GB RAM, gigabit Ethernet port and SATA. And it uses a compatible mechanical layout.
The better pi is already here…
An electronics design engineer’s perspective. Before doing a redesign consider the reasons for its success.
Linux based single board computers existed before the Pi – at reassuringly expensive prices. The Pi’s initial success was down to its price, but its continuing success is because it’s become a community supported standard. To build on that support and to avoid isolating existing users a replacement should break compatibility as little as possible.
In addition to the mechanical design fixes suggested by MarkC I’d fix the SD card connector and the dreadful audio output. I’d upgrade the network to gigabit ethernet and add any performance gains which Moore’s Law might have delivered in the interim.
Lastly I’d fix the power problems which do so much to mar the experience of beginners. It needs an alternative power connector with a matching, approved “wall wart” so that USB devices can be hot plugged without crashing the thing. This would also ensure that the Pi meets the USB power output spec on its ports. Then you wouldn’t need to add a powered hub to virtually every project.
It needs a host of detailed engineering fixes to make it the board it might have been, while maintaining as much software compatibility as possible.
The rPi is a teaching tool first and foremost and to that extent should not be altered in a way that the current stock are all superseded.
Yes a faster processor and more memory would be nice but any replacement must be backward compatible.
Given that the most common OS run on the Pi is Linux worries about “breaking compatibly” with a hardware change I think is being overstated. Userspace will remain compatible, if they keep the same GPIO headers etc then any shields will remain compatible. Building a kernel with the right modules to run on 2 arm based boards is not that difficult, so any distro could support both.
In terms of changes I would like to see :-
Fix the usb problems
Sata port
Maybe More Ram, Faster CPU (not that important)
PoE would be nice (not essential)
Since they brought out the dimm pi for industry, could they make a module board for adding multiple dimm pi together? Akin to the new arm servers with module cpus. That way they keep the pi technically the same, but also allows for people who need more power to add those modules. Though no the most cost effective way.
Yes, even at a higher price…a more powerful Pi would be really nice to have
Maybe a dual core 1Ghz, 1GB RAM, even a SOC compatible with Ubuntu
People buying the more expensive model would be helping subsidize the cheaper model for those who are happy with the existing model. I, for one, wouldn’t mind paying more for a faster / more powerful Pi
People who do, make do; people who who won’t, want more. Specifications could always be a little better…as I always say:-
It’s not having what you want
It’s wanting what you’ve got
(I’m gonna soak up the sun
I’m gonna tell everyone
To lighten up )
Go here to find a survey that gives you potential to win a Minnowboard Max.
https://01.org/iot/idfchina. Hope the winner is a LV reader! ONLY ONE MORE DAY
No, it isn’t. Banana Pi does have the RAM, better processor, Giga and SATA, it’s true. But the developers have been lazy with the development of the operating system. I have one, tried it with the Cubieboard OS (didn’t even boot) then tried it with Raspbian. Raspbian worked, yes, but a lot of the functionality doesn’t (yet) work such as the GPIO. And it’s not mechanically compatible, regardless of what they say – power port is in a different place and the spacing of the rest of the ports is off by a significant margin.
This should have appeared as a reply to a comment…