Suggestions for version 3

[johnbanks]

My biggest problem with the Pi are the SD cards.

I get errors out of nowhere far too often.

Just yesterday, working on a project. Walking an SD card between a model A and model B. Always being careful.

Then out of nowhere the card has tons of errors, have to stop everything and make a new card.

It really disappoints me that I can't trust that the Pi is going to be reliable no matter how careful I am.

That's when I start wondering when Arduino or Beagle are going to drop a nice HD video chip on one of their new Linux boards and I would have to leave the Pi behind.

I just greatly dislike the 'works fine! oh, now the card is trashed' experiences I've had with the Pi.

Really wish this could be locked down with some trick. Like the write protect switch on the cards actually working on the Pi. Blocking a pin connection between the card and pi? Maybe a never-write option in the config. Maybe a place for users to solder in some memory where the Pi can boot from instead.

But the best to me would be to drop the SD card for something much more stable.

There are many posts about running from a USB stick and only just booting from the SDHC which many people claim solves/minimizes the corruption problem.
I wonder whether the option to boot from a USB stick could somehow be be hard coded in - test if an SDHC is present and if not attempt to boot from USB - but I am no expert

[Lob0426]

Sorry if you felt I was dissing the thread, I was just pointing out that a higher price point PI IMO is not required all the while the Broadcom SoC is available..


My suggestions for future PI = not needed (apart from more memory which is not feasible)

Reason for not needed = other boards available which already have the the must-have features

If complete with other boards = fragmentation of the existing Pi Community


Regurgitation: add-on boards bring the Pi to the price point of the competitors, therefore one can only pay for what they need !


I will now go away and lie down in a darkened room with the Magic Mushrooms

Maybe buy this ??? http://www.raspberrypi.org/phpBB3/viewt ... 63&t=53875

I also do not think the current RasPi needs much of anything. I do think a future version for advanced students might be needed.

I think the retail price of add on boards is why people want to see things added to the RasPi. The real problem is that everyone needs something different. So an item that works for one, adds unnecessary cost for others.

I saw that iteastudio earlier. It is kinda of unclear that you do get the core board and the base. Looks interesting none the less. Priced a bit high in my opinion.

What I would like to see is a RasPi cased inside a keyboard for education. It seems that would be the best plan. I have been wondering why with all the cases that are out there, no company has a keyboard/case for sale. A number of them have been built but none on the market. Room for an internal hub with external ports would be a bonus.

Anyone know of a soft keyboard with a mouse pad available?

[Lob0426]

There are many posts about running from a USB stick and only just booting from the SDHC which many people claim solves/minimizes the corruption problem.
I wonder whether the option to boot from a USB stick could somehow be be hard coded in - test if an SDHC is present and if not attempt to boot from USB - but I am no expert

The boot sequence comes from the GPU. It cannot be changed. You can have just /boot on the SD card and then the /rootfs on a USB device. You only need like an old 32MB SD card to keep the boot on.

There are a number of posts and tutorials covering how to set this up.

The real question is why he is having those problems with his SD cards. I do not have those problems. Maybe static discharges is causing the problems for him.

[rpdom]

There are many posts about running from a USB stick and only just booting from the SDHC which many people claim solves/minimizes the corruption problem.
I wonder whether the option to boot from a USB stick could somehow be be hard coded in - test if an SDHC is present and if not attempt to boot from USB - but I am no expert

No. The Broadcom chip can only boot from SD card or special software on a directly connected USB host. As the model B has a USB hub chip, the USB option can't be used. It is possibly on a model A, as it doesn't have that hub chip, but you need special software to make it work and you won't be able to use any other USB devices like keyboards, mouse, storage etc... just not practical.

[W. H. Heydt]

Faster Processor, More memory, maybe more ports, including more GPIO pins?

Yes all are available in an alternative product, but not for £25.00, try £150++++

Actually...try $50. !GHz, 1GB RAM, more GPIO pins (96, I think), line in, SATA, otg USB (in addtion to 2 USB). Just hunt around a bit.

[Heater]

One thing we can learn from all of this is that the Pi must be really good.
Everyone who has tried some Pi wants more:)

[Heater]

W. H. Heydt,

Actually...try $50. !GHz, 1GB RAM, more GPIO pins (96, I think), line in, SATA, otg USB (in addtion to 2 USB). Just hunt around a bit.

Please name that board and tell me where I can get one at that price.

But really is that the point? If you need a gig of RAM or 96 I/O pins or SATA or whatever you can get it from many places. Whatever you need to pay you will pay.

[AndrewdAzotus]

WebPi,

I'd love a dual core chip with twice as much RAM.

So would I. Or more of all of that...

I know that would cost quite a bit more,

Yes of course. So it does not fit the aims of the Pi Foundation. I.E. Cheap computers for kids to play with.

but I think there's quite a lot of demand for devices with dual/quad core ARM chips.

Sure there is. And there are companies out there supplying such things, if you want to pay more.
The Rasperry Pi foundation is not a charity dedicated to to providing the latest and greatest and ever more powerful computer boards for everyone.

The Utilite has 1/2/4 cores and more RAM, but it's more money.

The only thing I would ask for on the next Pi is power and ethernet on the same edge of the board. Everything else (faster / more ram / more cores etc.) will come in it's own good time as costs allow.

[Lob0426]

W. H. Heydt,

Actually...try $50. !GHz, 1GB RAM, more GPIO pins (96, I think), line in, SATA, otg USB (in addtion to 2 USB). Just hunt around a bit.

Please name that board and tell me where I can get one at that price.

But really is that the point? If you need a gig of RAM or 96 I/O pins or SATA or whatever you can get it from many places. Whatever you need to pay you will pay.

I think he means the Cubieboard http://cubieboard.org/

[Oakham]

W. H. Heydt,

Actually...try $50. !GHz, 1GB RAM, more GPIO pins (96, I think), line in, SATA, otg USB (in addtion to 2 USB). Just hunt around a bit.

Please name that board and tell me where I can get one at that price.

But really is that the point? If you need a gig of RAM or 96 I/O pins or SATA or whatever you can get it from many places. Whatever you need to pay you will pay.

?? http://www.raspberrypi.org/phpBB3/viewt ... 62&t=53875 ??

[Heater]

Yep, turns out there are Pi competitors. Better features, similar price. If you can get them in your hand at that price.

So whats the problem?

Want a Pi get a Pi. Want something else, get something else?

Easy.

[AndrewdAzotus]

Having read this thread and seen the requests for faster CPU and more cores I wonder why.

If that's what you really want, buy a different computer or put one together from scrap. I've just built a dual core celeron with 1GB RAM and a 40GB HDD to run Debian from old computers that other people were throwing away so this was cheaper than my Pi to buy, but not cheaper to run! (I was also given a 33Mhz beastie running Win 3.1 - can't do much with that though other than smile about when Win 3.1 was thought of as "cutting edge")

On the Raspberry Pi, if it's not running fast enough, re-write your code to make it more efficient and, therefore, run faster.

Move on from Python and learn C. Once you've mastered C if it's still not fast enough, learn assembler and code stuff 'properly'.

I remain astonished that it can do so much and it cost so little. My Pi has (finally) brought me to Linux and it will (in the not too distant future, I hope) 'force' me to buy a soldering iron and some resistors and so on and start doing some hardware stuff.

What I fancy doing at the moment is buying 8 new Pi, having them all mounted in a picture frame on the wall (behind glass), all running and all doing something. I think there's even a pretty LED thingy to plug into the GPIO so I can program lots of pretty moving patterns.

Why? Because the RPF and the Raspberry Pi makes this all possible

[gritz]

Yep, turns out there are Pi competitors. Better features, similar price. If you can get them in your hand at that price.

So whats the problem?

Want a Pi get a Pi. Want something else, get something else?

Easy.

Welcome to humanity and humans that value prestige, branding and whatnot. Many of the folks who submit requests and suggestions are people that want something different to the current Pi, but want to have "Pi" written on it. It's not just the branding of course - there are other motivators, like "staying with what you know", remaining within the same community, a desire to see the product evolve / change direction / eject flying monkeys from it's USB ports, etc...

It's only when feature requests dry up that one should start to worry!

[OtherCrashOverride]

Having read this thread and seen the requests for faster CPU and more cores I wonder why.

ARMv7 for compatibility with the rest of the world: no more need to patch and port software designed for ARM devices (such as Ubuntu)
ARMv7 for NEON SIMD: enables new uses such as image processing, dsp, codecs, simulations
Faster: it would be nice to have a usable* www browser.

* this is subjective and while there are browsers that work, they are not something that you would want to use as a primary tool.

[OtherCrashOverride]

As for the discussion on price: Its a bit misleading to say the cost of a PI is $35.

For $35, the board will do absolutely nothing but collect dust. You have to add a SD card and power supply at the least and most will want a display of some sort. This means that the Total Cost of Ownership (TCO) is much steeper. So a $45 devices that eliminates the need for some of the extras is still quite a value proposition. Personally, I have found its much cheaper to use budget Android devices for hobby projects than to outfit a PI suitable to the same task.

[AndrewdAzotus]

Having read this thread and seen the requests for faster CPU and more cores I wonder why.

ARMv7 for compatibility with the rest of the world: no more need to patch and port software designed for ARM devices (such as Ubuntu)
ARMv7 for NEON SIMD: enables new uses such as image processing, dsp, codecs, simulations
Faster: it would be nice to have a usable* www browser.

* this is subjective and while there are browsers that work, they are not something that you would want to use as a primary tool.

where you suggest "faster = faster useable* browser". Yes, I do agree with you here. It is this reason that I don't use my Pi more and still use my laptop. Although I prefer Iceweasel to Midori.

And, that's a fair point about ARMv7 compatibility but it's a good argument for compatibility not necessarily for faster, it just so happens that the ARMv7 is faster too. I was just thinking back to the days when we programmed in BBC Basic on the BBC Micro and if that wasn't fast enough we switched to Assembler. There was no "I want a fast computer or more processors" - although the 6502c 2nd processor was fun.

[abishur]

My biggest problem with the Pi are the SD cards.

I get errors out of nowhere far too often.

Just yesterday, working on a project. Walking an SD card between a model A and model B. Always being careful.

Then out of nowhere the card has tons of errors, have to stop everything and make a new card.

It really disappoints me that I can't trust that the Pi is going to be reliable no matter how careful I am.

That's when I start wondering when Arduino or Beagle are going to drop a nice HD video chip on one of their new Linux boards and I would have to leave the Pi behind.

I just greatly dislike the 'works fine! oh, now the card is trashed' experiences I've had with the Pi.

Really wish this could be locked down with some trick. Like the write protect switch on the cards actually working on the Pi. Blocking a pin connection between the card and pi? Maybe a never-write option in the config. Maybe a place for users to solder in some memory where the Pi can boot from instead.

But the best to me would be to drop the SD card for something much more stable.

From my own testing I've found that this typically happens with overclocking. Specifically overclocking with a little overvoltage. There has been some speculation that a very small capacitor between the SD slot's ground and another pin (sorry I forget which off the top of my head, but it's been discussed on the forum) fixes this issue. I haven't been able to test that aspect of it, but I have fixed it by limiting myself to the overclock without overvoltage. Hopefully if we do see a Rev3 with the education release this and the USB port will be addressed, but even if they're not, while being annoying they're not show stoppers.

[Oakham]

And, that's a fair point about ARMv7 compatibility but it's a good argument for compatibility not necessarily for faster, it just so happens that the ARMv7 is faster too. I was just thinking back to the days when we programmed in BBC Basic on the BBC Micro and if that wasn't fast enough we switched to Assembler. There was no "I want a fast computer or more processors" - although the 6502c 2nd processor was fun.

Using an ARMv7 SoC would require a completely new product being designed from scratch....

[Oakham]

Hopefully if we do see a Rev3 with the education release this and the USB port will be addressed, but even if they're not, while being annoying they're not show stoppers.

Would not a USB GPIO expansion board overcome the USB issue ???

[AndrewdAzotus]

Using an ARMv7 SoC would require a completely new product being designed from scratch....

Is this really a problem? After all, the Raspberry Pi was a completely new product designed from scratch and see how we all love that!

[Oakham]

What about all the people who will now have an obsolete Pi with the ARMv6, will all those people go out and buy a completely new product which is not backward compatible ?

I appreciate that the ARMv7 would of been a better platform but 2+ years ago it would of not been cost-effective, the problem is that now there are so many other boards using ARMv7 that a nouveau Pi could lose it big community..........

Though maybe going ARMv7 will allow a standardisation of Linux Distributions across the various boards....

[OtherCrashOverride]

What about all the people who will now have an obsolete Pi with the ARMv6, will all those people go out and buy a completely new product which is not backward compatible ?

I would not cry about my $25/$35* computer becoming obsolete; especially if I could get the 'next great thing' at the same price.

* we are talking about replacing the board, not all the accessories.

[AndrewdAzotus]

Why would my current Pi become obsolete just because there is a new model with an ARMv7 SoC on it?

It will still run the released versions of Raspbian and Apache and everything else.

All the Linux commmands I know and love will still work.

My Model B Pi will not stop being used just because I have a new model...

[Oakham]

Why would my current Pi become obsolete just because there is a new model with an ARMv7 SoC on it?

It will still run the released versions of Raspbian and Apache and everything else.

All the Linux commmands I know and love will still work.

My Model B Pi will not stop being used just because I have a new model...

Would the community still want to support ARMv6 ?

I am playing devil's advocate as it is a bigger picture than just having a few more ports and modern SoC !

Anyway the Chinese are a coming with their cheap devices.........................................................................

[OtherCrashOverride]

Would the community still want to support ARMv6 ?

Trick question! Nobody currently wants to support ARMv6. There is just no choice in the matter for Pi.

Python will still be Python. C will still be C. I doubt there would be any major disruption overnight.