What do people want to do that the rpi currently cannot do??

[mung]

I always notice complaints about rpi not having X component when a new pi gets released.

I sometimes think people do not realise what they need or what they can do with what they have, so I thought I would ask this idle question.

What do people want to do that the rpi currently cannot do?

And propose what I think the SOC needs for the next generation:

I think that a really flexible on SOC IO processor is needed. Something like the beaglebone PRU but with USB 3.1/c and analog ADC possibly even DAC all routable to any given pin and integrated with DMA and realtime processors to handle any necessary IO without ARM or VC4/5 intervention. I get the feeling this is something that is probably very hard to get on one piece of silicon but once implemented it makes a multi purpose chip suitable for almost any application. I am sure smart TVs, settop boxes, home automation, and robotics are growing markets with huge scope. I feel there is advantage to having the same processor system doing everything as it reduces users learning overhead and simplifies interoperability.

Obviously not really anything to do with educational goals but maybe in some ways related to it?

Education and learning should be about how to get the best from life, as life becomes more linked to computers what do we need to learn about them and what can computers do to make life better (please forgive the above hippy bullshit, obviously the most important thing in life is money!). How can you make it cheaper?

[Pithagoros]

I think, at the very minimum the next Pi should have a 64 qubit Quantum computing capability.

[jamesh]

I think, at the very minimum the next Pi should have a 64 qubit Quantum computing capability.

Yup, it's got that*. And a sandwich maker, and free pony.

But to stay on topic.

From my reading of the forums, the majority of people want wireless, and possibly bluetooth, and faster ethernet, and more memory, Oh, and even faster processors. And the price mustn't go up. Of course, you can want all you want, but the Foundation still needs to stay in business, and that business is education. What it has and and when it arrives? Who knows.

And an X driver, so many people waiting for the results of the Anholt software effort (which should be a real leap, even for existing PI's),. There is a lot of scope for more speed even using the original Pi's. I'd want lots of effort to go into that.

Oh, I forgot the SATA port, lots of people want those.





* No it hasn't

[gregeric]

These threads aren't always pointless...
viewtopic.php?f=63&t=495#p3725
And no, I don't mean that Liz got her pony. The other one sort of came true...

[gordon77]

Overall not much, maybe just the facility to plug in a battery to have a controlled shutdown on power loss.

[karrika]

The things I have missed during 2015:
The Raspberry Pi 2 would benefit from
- a faster SPI port
- microphone input
- a HAT specs that allows you to stack add-ons

The faster SPI would allow us to use Raspberries within industry for distributed data processing. Today Ethernet is the only viable option.

I also work with theater productions and use Raspberry Pi 2 for controlling light and sound. For using the Raspberry in front of a paying audience I need stackable HATs. The Cirrus Logic Audio Card is ok. But in addition to that I need my DiscoHAT http://discohat.com. These should stack nicely.

I wrote in another forum about ways to make stacking reality. Also documenting the pinouts for different boards is important. Something like http://pinout.xyz should evolve into a tool by which you can find extension boards that go together.

[mung]

I am not asking what the board needs to have, I am asking what people want to do with it that they cannot do, I am hoping I may see something that makes sense, or that will help me understand the future.

It seems to me the rpi is currently vastly under utilised most people hardly use the power that is available to its real limits.

I suppose wifi maybe the only thing important that is currently missing if the rpi is a general purpose internet learning computer.

But when people say 'more memory/cpu/gpu/gigabit network' I think why do they need it what do they want to use the rpi for?

[gordon77]

I think you can try and predict the future, with differing levels if success, l am not sure about understanding the future.

[jamesh]

I am not asking what the board needs to have, I am asking what people want to do with it that they cannot do, I am hoping I may see something that makes sense, or that will help me understand the future.

It seems to me the rpi is currently vastly under utilised most people hardly use the power that is available to its real limits.

I suppose wifi maybe the only thing important that is currently missing if the rpi is a general purpose internet learning computer.

But when people say 'more memory/cpu/gpu/gigabit network' I think why do they need it what do they want to use the rpi for?

Sorry, misread you post, which is actually more interesting that the usual.

The Pi is so cheap, that you can afford for it to be underutilised, because a processor board that may be more suitable power wise is almost always more expensive! And I mean lower power. Exception is Arduino boards of course, they are very cheap, but don;t have the full Linux ecosystem.

Lots of embedded system don't need fast 3D, or HDMI output, or H264 encode/decode etc. But you effectively get those for practically free if all you want is an ARM workhorse.

[W. H. Heydt]

Oh, I forgot the SATA port, lots of people want those.

Really? I thought I was the only one that wants that. (Well...probably a few "fellow travelers" that are mostly silent.)

[Joe Schmoe]

I still maintain that the two things that I wish the Pi could do that it doesn't do are:

1. Play sites that use Flash (No, I'm not talking about YouTube).
2. Run Windows (x86) apps (at some level - obviously, it's never going to run current versions - Windows 10, cough, cough). Note that some people claim to have done this, via QEMU, but I've yet to get around to testing it. It would be nice if the Foundation (or some else on the "official" side) could validate these claims and make it generally available. Yes, I know I'm just spit-balling here...

Those are the two things that the Pi currently cannot do.

[blc]

Personally, there's a lot that I'd like to be able to do with the Pi that it can't currently do or - perhaps more accurately - is ill-suited to do, such as:

Gigabit ethernet, so that I can run a file or download server
SATA, to facilitate the above
USB3 bus, so that we can add features which would currently be pointless on a USB2 bus (e.g., a USB SSD filesystem)

However, all of those features would significantly increase price and push the Pi into the "low-cost x86" world. Maybe that's not a bad thing for a consumer, but it doesn't really sit well with the educational goals. Plus that market is already very well served, by both used and new hardware (and usually far cheaper than people realise).

I never expected the original Model B, complete with all its hiccups and quirks, to be as capable as it was. Every major advance we've had since then is a bonus IMO: more memory, more stable power, reduced power usage, more USB ports, a far more powerful quad-core model with double the memory, an ultra low-cost variant which costs about as much as a pint (in its basic form).

[W. H. Heydt]

Personally, there's a lot that I'd like to be able to do with the Pi that it can't currently do or - perhaps more accurately - is ill-suited to do, such as:

Gigabit ethernet, so that I can run a file or download server
SATA, to facilitate the above
USB3 bus, so that we can add features which would currently be pointless on a USB2 bus (e.g., a USB SSD filesystem)

However, all of those features would significantly increase price and push the Pi into the "low-cost x86" world. Maybe that's not a bad thing for a consumer, but it doesn't really sit well with the educational goals. Plus that market is already very well served, by both used and new hardware (and usually far cheaper than people realise).

I never expected the original Model B, complete with all its hiccups and quirks, to be as capable as it was. Every major advance we've had since then is a bonus IMO: more memory, more stable power, reduced power usage, more USB ports, a far more powerful quad-core model with double the memory, an ultra low-cost variant which costs about as much as a pint (in its basic form).

Gigabit Ethernet is really only going to work on a LAN, unless you are one of the lucky few in the world who has a broadband connection that runs well above 100Mb/s. Much better item for the "wish list" would be for the Ethernet port to be native to the SoC and not run through a USB connection.

SATA: My solution is to use Allwinner A20 boards to get that (and it's SATA II, so not *super* fast, but pretty darned good with an SSD).

USB 3: If we got SATA, that removes much of the desire for USB 3.

While some time in the distant future, all of those would be "nice to have", to get the practical effect, getting Ethernet off the USB (even if it's still 10/100) and adding a SATA connector would cover a *lot* of the use cases you're looking at.

The other item that I think should be on the list is an RTC. Even so, the need for that is minimal. For example, I just ordered 10 RTC modules that fit Pis (I have some in use already) at a cost of $2.58 each. I could see a built in RTC only if a new SoC is being designed anyway.

[blc]

Gigabit Ethernet is really only going to work on a LAN, unless you are one of the lucky few in the world who has a broadband connection that runs well above 100Mb/s. Much better item for the "wish list" would be for the Ethernet port to be native to the SoC and not run through a USB connection.

SATA: My solution is to use Allwinner A20 boards to get that (and it's SATA II, so not *super* fast, but pretty darned good with an SSD).

USB 3: If we got SATA, that removes much of the desire for USB 3.

While some time in the distant future, all of those would be "nice to have", to get the practical effect, getting Ethernet off the USB (even if it's still 10/100) and adding a SATA connector would cover a *lot* of the use cases you're looking at.

The other item that I think should be on the list is an RTC. Even so, the need for that is minimal. For example, I just ordered 10 RTC modules that fit Pis (I have some in use already) at a cost of $2.58 each. I could see a built in RTC only if a new SoC is being designed anyway.

I am one of the lucky few out there with an internet connection running at >100Mb/s.

In any case my point wasn't necessarily "I want the Pi to do X", more "these things on a Pi would be nice to have but they're already available on other hardware, so why bother when the cost of adding them would bring the Pi's price up in-line with the other hardware". For example: my fileserver used to be a Core2Duo-era PC - not exactly the fastest machine in the world, but more than adequate for the job it's doing. The original Pi Model B already ticked a lot of the boxes that the Foundation was aiming for: a general purpose educational computer cheap enough to almost be disposable (compared to the cost of equivalent SBCs or computers at the time). The fact that it can do so much more is bloody brilliant.

[rpdom]

I am one of the lucky few out there with an internet connection running at >100Mb/s.

My internet connection is only 50Mb/s, but that's because I chose the cheapest option my ISP provides and I really don't need faster

My internal LAN runs at 1Gb/s, but only a few devices have suitable interfaces for that (my ancient laptop being one of them). Still better than when I had 100Mb/s (or even 10)

[blachanc]

Oh, I forgot the SATA port, lots of people want those.

Really? I thought I was the only one that wants that. (Well...probably a few "fellow travelers" that are mostly silent.)

Did you see this (need compute module and no GPIO)?:

WD 2.5" SATA-to-Raspberry Pi Adapter Board

WD_sata2Pi.jpg (22.47 KiB) Viewed 5609 times

Ben

[stderr]

Gigabit ethernet, so that I can run a file or download server
SATA, to facilitate the above USB3 bus

Basically the big bottleneck problem needs fixing. But that bottleneck is one reason why the board is so cheap. Of course there are other boards in the range that are less so constrained.

[W. H. Heydt]

Oh, I forgot the SATA port, lots of people want those.

Really? I thought I was the only one that wants that. (Well...probably a few "fellow travelers" that are mostly silent.)

Did you see this (need compute module and no GPIO)?:

WD 2.5" SATA-to-Raspberry Pi Adapter Board

WD_sata2Pi.jpg

Ben

that's extremely interesting, and--no--I hadn't seen it. It could use an Ethernet port, though (and that can be handled with inexpensive add-on hardware). If/when the CM2 arrives, that would be a really sweet little system.

[quipu]

Raspberry Pi ALPHA with sata, rtc, gigabit ethernet, mic plug, better gpu, power and battery contol unit integrated.
I want that And I also want that cheap and small too

[TimG]

What do people want to do that the rpi currently cannot do?

That's a much more interesting question than the usual feature wishlist.

With the release of the Zero, the RPi is becoming even more attractive as a microcontroller replacement in small battery-powered projects. What I'd like to see is an RPi which can run for much longer on modest battery power. This could be handled via low-power idling and very low-power standby modes, entered and exited without the cumbersome business of shutdown and reboot.

On the software side I'd like to see improved support for read-only disks, allowing the Pi to better withstand abrupt power removal. A good old-fashioned on-off power switch is much more practical than a controlled shutdown for any project which doesn't include a keyboard.

[jbeale]

Many of my projects need hard-real-time performance. In theory I guess I could use a Pi, if I forget Linux and program on the bare metal, but that's more work than I want to do. For now I'm happy using a "Teensy" or other Arduino-type microcontroller board for those, the programming is simpler for me to understand than any kind of real-time kernel solution. Having a very low-power sleep mode (which you could resume from quickly) would certainly help for standalone battery projects.

[neurojab]

Well, if this is a pie in the sky wishlist, I'd like to be able to do the following:
* Edit raw photos and videos - I need at least 4GB of RAM for this.
* Play DRM video content (netflix and google play) AKA run android as dual-boot. Yes I know there are other boards that can do this.
* Build servers (SATA and gigabit ethernet)
* Deploy battery powered wifi enabled IOT devices. Yes this is possible today but could be more convenient.

That said, I love the pi2b pricing and specs. I'm not suggesting replacing it. It would be cool to also have a model S (Server) model with SATA and a model D (Desktop replacement) model with more RAM and a model W with an included wifi chip and possibly running on 3v with battery friendly connectors.

[Chris-The-Tuner]

Well i join in on this topic and just say what i would like to see (and why).

For the moment my only problems using the RPi are just the lack of Speed when it goes to data tranfser.
The RPi is cheap, and i would love to use it as kind of a NAS, but since the RPi does not have USB3.0 and gigabit ethernet this is kind of a problem.

I don't like the idea of pre-made NAS servers, i prefer my own system here.
I know using most Synology disk stations because i install them for my customers if they want me to, but as i said i prefer homemade solutions...

Because of the RPi not being able to deliver data fast, i switched to an ODROID-XU4 for the use as a fileserver.

A "normal" computer/server is not going to make it for me, i live in germany and i pay 0,28EUR/kWh...
I have already 3 Servers running 24/7 for Gameservers and i don't want to add up to this just to have a NAS for my Kodi media centers (RPi of course !).

Everything else on the RPi is perfect for me, i never missed something else...

[karrika]

There is actually one thing that could make a big difference for Raspberry Pi use. Most small boxes are vulnerable for all kind of malware attacks.

Perhaps it would be possible to create a Raspberry Pi with read-only SD filesystem. It would have enough RAM for mirroring the essential stuff in a RAM filesystem. For permanent storage we would use USB memory sticks.

So when this computer gets attacked you can detect it by some tripwire style monitoring and reboot the computer to get rid of the malware.

Perhaps you could have a mechanical switch to disable the writes?

[rpdom]

There is actually one thing that could make a big difference for Raspberry Pi use. Most small boxes are vulnerable for all kind of malware attacks.

Perhaps it would be possible to create a Raspberry Pi with read-only SD filesystem. It would have enough RAM for mirroring the essential stuff in a RAM filesystem. For permanent storage we would use USB memory sticks.

So when this computer gets attacked you can detect it by some tripwire style monitoring and reboot the computer to get rid of the malware.

Perhaps you could have a mechanical switch to disable the writes?

It's been done many times (not with the switch though). Although adding any sort of permanent storage pretty much defeats the object of the exercise.