User avatar
Burngate
Posts: 5884
Joined: Thu Sep 29, 2011 4:34 pm
Location: Berkshire UK Tralfamadore
Contact: Website

Re: is the pi worth it?

Wed May 15, 2019 6:25 pm

Just a quick thought, about this Backwards compatibility / Scared of the future / landfill business.
Front page on Monday, picture of Soyuz, taken on 1B+ & camera.

If we're thinking of losing backwards compatability, then anything written for post-3B+ won't work on the ISS.
And landfill is a long way down.

hippy
Posts: 5345
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: is the pi worth it?

Wed May 15, 2019 7:10 pm

Burngate wrote:
Wed May 15, 2019 6:25 pm
If we're thinking of losing backwards compatability, then anything written for post-3B+ won't work on the ISS.
It depends on what this 'backwards compatibility' amounts to. I would expect Python code to run just as well on a Pi 1B as it would on a post-Pi 4B. So long as any new code doesn't depend on something which isn't supported by the 1B. And I would expect that to apply to any code, though one might have to recompile it for a 1B.

There will be some issues if there were a loss of backwards compatibility but I would say "anything written for post-3B+ won't work" is stretching it.

jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 22697
Joined: Sat Jul 30, 2011 7:41 pm

Re: is the pi worth it?

Wed May 15, 2019 8:15 pm

hippy wrote:
Wed May 15, 2019 7:10 pm
Burngate wrote:
Wed May 15, 2019 6:25 pm
If we're thinking of losing backwards compatability, then anything written for post-3B+ won't work on the ISS.
It depends on what this 'backwards compatibility' amounts to. I would expect Python code to run just as well on a Pi 1B as it would on a post-Pi 4B. So long as any new code doesn't depend on something which isn't supported by the 1B. And I would expect that to apply to any code, though one might have to recompile it for a 1B.

There will be some issues if there were a loss of backwards compatibility but I would say "anything written for post-3B+ won't work" is stretching it.
GPIO, Camera, Graphics, any software that talks to HW etc. We of course are moving to MESA for graphics (Anholt driver) which should make us much more forwards compatible as well. Of course, a lot of stuff will "just work", but we need to ensure HW changes don't cause problems with existing software,or at leastmake any problems easy to fix.
Principal Software Engineer at Raspberry Pi (Trading) Ltd.
Contrary to popular belief, humorous signatures are allowed. Here's an example...
"My grief counseller just died, luckily, he was so good, I didn't care."

hippy
Posts: 5345
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: is the pi worth it?

Thu May 16, 2019 1:05 am

jamesh wrote:
Wed May 15, 2019 8:15 pm
GPIO, Camera, Graphics, any software that talks to HW etc. ... Of course, a lot of stuff will "just work", but we need to ensure HW changes don't cause problems with existing software,or at leastmake any problems easy to fix.
Absolutely, but I would expect that to be abstracted and hidden by system services, tools and libraries so, while a lot of work for those providing and implementing those parts, for the average user nothing much changes. Just like moving programs between platforms now, it's mostly a case of recompilation and it works.

GPIO I would think would be the most problematic area for users as direct access to the I/O space seems to have been the recommended way of doing things. But I imagine that could be resolved by exposing some memory as that I/O space, trapping reads and write, and having a HAL sit between that and real hardware.

That potential issue might be a good motivation to start moving to a better, managed, interface to GPIO and peripherals, which would also have advantages in itself.

User avatar
Gavinmc42
Posts: 3150
Joined: Wed Aug 28, 2013 3:31 am

Re: is the pi worth it?

Thu May 16, 2019 2:21 am

GPIO, Camera, Graphics, any software that talks to HW etc
Been thinking about some more.
RPF probably has more exposure to industry usage where this can be very important.
For the home, hobbyist, school users the hardware is less an issue as any changes to hardware might only need a software change.
Of those, only the School stuff is the RPF's stated Mission, everyone else is just a bonus.

How many Cameras/LCDs the embedded industry uses is not known to most of us.
I can see industry might get upset if there is any side effects of hardware changes;)

Will I lose the OpenVG hardware but gain Vulkan capability?
One is supposedly dead and the other is the future.
About the only ones I can think of who maybe using OpenVG are the QT guys.

Is there going to be a time when the Pi stops being worth it?
The first interesting RISC-V boards are starting to appear but they will take a few more years to get to volume and there is still a lack of speed/GPU.
Any RISC-V/GPU had better be Vulkan hardware capable as that is where the software tools are at now.
Backwards capability is not needed for that market. Will RISC-V threaten the Pi market?
Probably not for at least 5 years and then RPF will perhaps be up to the Pi6?

Will the Pi6 be worth it? We will need to wait and see:lol:
It will be very hard for anyone to displace the Pi's School/Education Mission.
China and India would have to product their own Education hardware and software products.
How big would RPF have to be to make an impact in those markets?
Only greenhouse warming or asteroid impact could stop the snowball Eben has started?
I'm dancing on Rainbows.
Raspberries are not Apples or Oranges

hippy
Posts: 5345
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: is the pi worth it?

Thu May 16, 2019 12:37 pm

Gavinmc42 wrote:
Thu May 16, 2019 2:21 am
GPIO, Camera, Graphics, any software that talks to HW etc
Been thinking about some more.
RPF probably has more exposure to industry usage where this can be very important.
For the home, hobbyist, school users the hardware is less an issue as any changes to hardware might only need a software change.
Everyone should know by now that, if one wants to be able to cater for hardware or platform changes, one doesn't access hardware directly, but through libraries and modules, which provide the means to cater for change and add abstraction layers should that become necessary..

If people haven't learned that they are largely the creators of the problems they face when hardware and platform changes occur.
Gavinmc42 wrote:
Thu May 16, 2019 2:21 am
Is there going to be a time when the Pi stops being worth it?
If the Pi cannot maintain its standing alongside other things which come in the future, through backwards compatibility issues or for other reasons, then such a time could come. They will have boxed themselves into a corner which they cannot get out of.

But I don't believe it will. As I have said, I don't believe backwards compatibility will be seen as so sacrosanct that they would rather let the ship slowly sink than change course.

Or, more likely, they will find a way to maintain backwards compatibility, even as the platform and hardware changes.

stuartiannaylor

Re: is the pi worth it?

Thu May 16, 2019 1:17 pm

Just talking about perfect Pi4 and what should be on a PI and what shouldn't is interesting and just to say why I mentioned my perfect Pi is devoid of everything apart from 3x M.2 slots and a 40 pi GPIO.

If you look at the China clones there is a whole range of various prices and often the difference is board price due to size and accommodation of a plethora of ports and connectors which greatly add to the overall build.
Any one singular instance practically always uses a tiny fraction of the ports and connectors available, most is redundant in single application but all useful for different use.

That is why I think scrap the lot and have a new high-speed GPIO and M.2 was just the first thing that I could think of that would be applicable.
Maybe 1x M.2, maybe 2x M.2 as then there is no limits to ports and connectors as you buy a base Pi and then get the M.2 daughter boards you need for application.

That one simple addition to a more beefy Pi could actually provide far more for less as M.2 just think of as PCIE x4 lanes so anything from storage to AI accelerator or even external GPU can sit on that bus.
The power of the Pi4 will offer a host of possibilities but if its the same format as the original in connectors, I don't think its possible to do it justice.

Raspberry are in a unique position where they could set a new form factor with a new high speed GPIO that would standardise this arena and have great benefits for education that standardisation brings.

Jamesh just think if a Pi4 had a M.2 I would have to stop moaning about VC5 as there would prob be a rake of cheap and cheerful external GPUs :)

The RK3399 has a single 4 lane PCIe available but also provides a host of other connectors as it was essentially of set top / phone origin that started to take an application from.
I am sort of expecting a Pi4 being somewhere in between RK3328 & RK3399 level maybe even out perform it, but raspberry have a unique perspective where they can sell in mass bulk general purpose application SoC systems.

Why I think its time for Raspberry to completely rethink form factor and application as what ever they make it will sell.
The backward compatibility argument is a non starter as maybe we should all still have ISA slots in our PCs.

A Pi4 should be able to do what earlier models did and may need a daughter board but Linux and the Kernel is the abstraction layer for software compatibility not hardware.
An older model should not be able to do what a newer model does like all older models that just lack the capability.

Once more reading in between the lines of the "unofficial raspberry" position by "official raspberry" members I am sat here thinking "Oh no raspberry" your in a position to do it again and become even more massive as you own the namespace of massive value.
If what we get is just incremental and backward compatible, forced to use a certain element of intuition of reading in between very few lines and info avail, some of us with a bit of imagination are slightly downhearted to what in time the adoption of conservatism to the loss of a fairly modest ambition will mean.

I said M.2 as an example so I could say get rid of much of the now defunct Pi legacy, could be any but can not think of another high speed bus that is as available or inoperable from connectors to use as M.2 currently.
The Pi4 should have less but be more modular and would love to see Pi4 M.2 clusters with what some crazy ass developer could do and a whole rake of next gen tinkering that Raspberry and Pi is an absolutely brilliant and cheap platform to use.

From what I hear, what I read in all honesty I am thinking its sounding like a huge opportunity is going to be missed for us and Raspberry themselves.
That just me and last I will say, which is a firm promise but just one thing I want to say is "high speed bus" as after that don't really care what Pi4 has because I will be happy and know it has a strong new future.

hippy
Posts: 5345
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: is the pi worth it?

Thu May 16, 2019 1:58 pm

stuartiannaylor wrote:
Thu May 16, 2019 1:17 pm
Raspberry are in a unique position where they could set a new form factor with a new high speed GPIO that would standardise this arena and have great benefits for education that standardisation brings.
That's probably more appealing to the ears of RPT than "I think it would be great if the Pi-5 had this". It could put them ahead of the competition and have them playing catch-up. But I doubt they wouldn't would want to kill the 40-way GPIO market so it would likely need to be additional.

This is where "backwards compatibility" is important. It's okay to ask users to suffer some pain, but very different when one drives them away, alienates them, has them choosing to go elsewhere to continue with what they bought into rather than go with the 'new shiny' they are expected to embrace. Loose too many and one may not easily get back to where one was.
Last edited by hippy on Thu May 16, 2019 2:01 pm, edited 2 times in total.

stuartiannaylor

Re: is the pi worth it?

Thu May 16, 2019 2:00 pm

hippy wrote:
Thu May 16, 2019 1:58 pm
stuartiannaylor wrote:
Thu May 16, 2019 1:17 pm
Raspberry are in a unique position where they could set a new form factor with a new high speed GPIO that would standardise this arena and have great benefits for education that standardisation brings.
That's probably more appealing to the ears of RPT than "I think it would be great if the Pi-5 had this". It could put them ahead of the competition and have them playing catch-up. But I doubt they wouldn't want to kill the 40-way GPIO market so it would likely need to be additional.

This is where "backwards compatibility" is important. It's okay to ask users to suffer some pain, but very different when one drives them away, alienates them, has them choosing to go elsewhere to continue with what they bought into rather than go with the 'new shiny' they are expected to embrace. Loose too many and one may not easily get back to where one was.
No 40 pin GPIO is ubiquitous and wouldn't be without it, but what shape is the raspberries high speed bus to be the faster accompanying brother gpio?
But the rest when you do could be omitted as so much could be provided by daughter cards but to set that new format that is the Pi, its up there with the IBM PC in form factor standardisation and has huge benefits for everyone and it is a sword they should wield like a mighty raspberry knight :)

But hey enough of that for me as said enough you get my gist and from the sounds of things a bit worried long term but hey its not my call, but just saying.

jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 22697
Joined: Sat Jul 30, 2011 7:41 pm

Re: is the pi worth it?

Thu May 16, 2019 2:12 pm

stuartiannaylor wrote:
Thu May 16, 2019 1:17 pm
Just talking about perfect Pi4 and what should be on a PI and what shouldn't is interesting and just to say why I mentioned my perfect Pi is devoid of everything apart from 3x M.2 slots and a 40 pi GPIO.

If you look at the China clones there is a whole range of various prices and often the difference is board price due to size and accommodation of a plethora of ports and connectors which greatly add to the overall build.
Any one singular instance practically always uses a tiny fraction of the ports and connectors available, most is redundant in single application but all useful for different use.

That is why I think scrap the lot and have a new high-speed GPIO and M.2 was just the first thing that I could think of that would be applicable.
Maybe 1x M.2, maybe 2x M.2 as then there is no limits to ports and connectors as you buy a base Pi and then get the M.2 daughter boards you need for application.

That one simple addition to a more beefy Pi could actually provide far more for less as M.2 just think of as PCIE x4 lanes so anything from storage to AI accelerator or even external GPU can sit on that bus.
The power of the Pi4 will offer a host of possibilities but if its the same format as the original in connectors, I don't think its possible to do it justice.

Raspberry are in a unique position where they could set a new form factor with a new high speed GPIO that would standardise this arena and have great benefits for education that standardisation brings.

Jamesh just think if a Pi4 had a M.2 I would have to stop moaning about VC5 as there would prob be a rake of cheap and cheerful external GPUs :)

The RK3399 has a single 4 lane PCIe available but also provides a host of other connectors as it was essentially of set top / phone origin that started to take an application from.
I am sort of expecting a Pi4 being somewhere in between RK3328 & RK3399 level maybe even out perform it, but raspberry have a unique perspective where they can sell in mass bulk general purpose application SoC systems.

Why I think its time for Raspberry to completely rethink form factor and application as what ever they make it will sell.
The backward compatibility argument is a non starter as maybe we should all still have ISA slots in our PCs.

A Pi4 should be able to do what earlier models did and may need a daughter board but Linux and the Kernel is the abstraction layer for software compatibility not hardware.
An older model should not be able to do what a newer model does like all older models that just lack the capability.

Once more reading in between the lines of the "unofficial raspberry" position by "official raspberry" members I am sat here thinking "Oh no raspberry" your in a position to do it again and become even more massive as you own the namespace of massive value.
If what we get is just incremental and backward compatible, forced to use a certain element of intuition of reading in between very few lines and info avail, some of us with a bit of imagination are slightly downhearted to what in time the adoption of conservatism to the loss of a fairly modest ambition will mean.

I said M.2 as an example so I could say get rid of much of the now defunct Pi legacy, could be any but can not think of another high speed bus that is as available or inoperable from connectors to use as M.2 currently.
The Pi4 should have less but be more modular and would love to see Pi4 M.2 clusters with what some crazy ass developer could do and a whole rake of next gen tinkering that Raspberry and Pi is an absolutely brilliant and cheap platform to use.

From what I hear, what I read in all honesty I am thinking its sounding like a huge opportunity is going to be missed for us and Raspberry themselves.
That just me and last I will say, which is a firm promise but just one thing I want to say is "high speed bus" as after that don't really care what Pi4 has because I will be happy and know it has a strong new future.
More concern trolling. I did ask you to stop. I have other, more effective ways of asking. Just for your information, there is nothing 'defunct' about the Pi format.

But I will answer. You are talking about designing an entirely different device. Not a Raspberry Pi. If we move away from the current form factor, its not a Raspberry Pi, and we lose. Feel free to design your own board, and see if it sells. I doubt it. You can design what ever you like into it, as the form factor is completely different, it's doesn't need to use the BCM range of SoC, doesn't need VC4 etc. You will need to start from scratch with software though, cannot leverage any of the existing stuff.

However, if the format was thought useful, don't you think others would already have done it and succeeded?

And as I said before, the Pi4 spec is already set in stone. Even the Pi5 has a lot of spec already defined. We were talking possibilities for Pi6 over lunch yesterday! But that is many years away so very blue sky.

As for the Pi4, if we don't sell 7-10 million in the first year, I would be surprised. Then will you shut up?
Principal Software Engineer at Raspberry Pi (Trading) Ltd.
Contrary to popular belief, humorous signatures are allowed. Here's an example...
"My grief counseller just died, luckily, he was so good, I didn't care."

Heater
Posts: 12639
Joined: Tue Jul 17, 2012 3:02 pm

Re: is the pi worth it?

Thu May 16, 2019 2:17 pm

stuartiannaylor,

I don't know about any future Pi but thanks for putting forward the M.2 connectors. They might be just what I need for some project PCB's I'm thinking about. Never mind the defined pin outs or electrical/protocol spec. It makes for a small cheap, compact, commonly available connector.
The backward compatibility argument is a non starter as maybe we should all still have ISA slots in our PCs.
Yes please. It used to be so easy to lash up a circuit on a proto board and plug it into the back plane.

W. H. Heydt
Posts: 10301
Joined: Fri Mar 09, 2012 7:36 pm
Location: Vallejo, CA (US)

Re: is the pi worth it?

Thu May 16, 2019 2:36 pm

jamesh wrote:
Thu May 16, 2019 2:12 pm
As for the Pi4, if we don't sell 7-10 million in the first year, I would be surprised. Then will you shut up?
Probably not. Just take the same text and do a global search and replace of "Pi4" with "Pi5".

(FYI...my experience and observation is that nearly all people with wish lists for future products, be they hardware or software, seriously underestimate the lead times for new features.)

Andyroo
Posts: 3306
Joined: Sat Jun 16, 2018 12:49 am
Location: Lincs U.K.

Re: is the pi worth it?

Thu May 16, 2019 2:43 pm

@jamesh - You will never win.

You folk must have the toughest skin of the lot of us. My hat is off to you.

Can we have a Pi '-1' please - low power, limited functionality, RTOS - I'll shut up now as I know better :lol:
Need Pi spray - these things are breeding in my house...

stuartiannaylor

Re: is the pi worth it?

Thu May 16, 2019 2:57 pm

Andyroo wrote:
Thu May 16, 2019 2:43 pm
@jamesh - You will never win.

You folk must have the toughest skin of the lot of us. My hat is off to you.

Can we have a Pi '-1' please - low power, limited functionality, RTOS - I'll shut up now as I know better :lol:
Called a pi-zero mate you should get one they r great

ejolson
Posts: 3039
Joined: Tue Mar 18, 2014 11:47 am

Re: is the pi worth it?

Thu May 16, 2019 2:59 pm

jamesh wrote:
Thu May 16, 2019 2:12 pm
However, if the format was thought useful, don't you think others would already have done it and succeeded?
When the Pi first came out it was pretty clear that the format was useful and that others had not already done it. That's why millions sold rather than thousands.

Using m.2 instead of USB3 sounds a bit like when IBM replaced the reverse-compatible AT bus with micro-channel. Computer history is full of examples where unwise technology decisions sank a project more thoroughly than the unwise marketing decisions, though such things usually went together.

At some point, it can help sell the current product when there is a known roadmap for the next that guarantees an evolutionary future with reverse compatibility. For example, IBM would have a lot harder time selling Power9 if it wasn't known that Power10 was coming and key details of the planned improvements. Purchases of DEC Alpha were drastically curtailed when it was understood that Compaq had stopped future development; sales of SPARC are currently experiencing a similar crisis.

While a Pi 4 in development is reassuring, a product roadmap phrased in terms of sandwich toasters and small equestrian animals is not something that can be taken to a board meeting and used to convince people who believe in Apple and Dell that the makers of Pi have long-term plans for the future.

While Raspberry Pi is good, technology advisors for schools may continue to recommend ipads and PCs, because an uphill battle with no visible roadmap makes them question is the Pi worth it?
Last edited by ejolson on Thu May 16, 2019 3:22 pm, edited 1 time in total.

Andyroo
Posts: 3306
Joined: Sat Jun 16, 2018 12:49 am
Location: Lincs U.K.

Re: is the pi worth it?

Thu May 16, 2019 3:22 pm

ejolson wrote:
Thu May 16, 2019 2:59 pm
...
While being told there is a Pi 4 in development is reassuring, a product roadmap phrased in terms of sandwich toasters and small equestrian animals is not something that can be taken to a board meeting and used to convince people who believe in Apple and Dell that the makers of Pi have long-term plans for the future.

While Raspberry Pi is good, technology advisors for schools may continue to recommend ipads and PCs, because an uphill battle with no visible roadmap makes them question is the Pi worth it?
Have to disagree - totally different market and often totally different budgets. Commercially buying a Pi or two would not even need a capital expenditure form (my entry point was £1K) but buying one PC / Laptop would so the board would not be interested (asset register - the bane of It managers). Same with long term support - these would be throw away items or just dedicated to one job and any other use is a bonus.

In the classroom, the Pi would be seen as a teaching aid and tied to a course - as long as it hit that years exam needs it gets in (no choice exams rule). If the budget gets tight then the school becomes an academy and gets new funding (bitter - me - not) or looks to use the Pi in two or three subjects and split the costs.

The cost of the Pi becomes an issue when you add the external components to it (sensors / boards / passives etc) for 20 or so work sets - then folk notice BUT you could not do the low level STEM work with a PC. I wonder if buying a kit is cheaper than training your techs up to order bits and make the kits (hum thats a thought)... Thank goodness the UK is again waking up to using technology rather than 'this is how you set the margins in Word'.

I've know places 'sneak' these in a printer toner, stationary and petty cash - its better for IT to work with them rather than try to stop them by paperwork / board discussions.
Need Pi spray - these things are breeding in my house...

ejolson
Posts: 3039
Joined: Tue Mar 18, 2014 11:47 am

Re: is the pi worth it?

Thu May 16, 2019 3:32 pm

Andyroo wrote:
Thu May 16, 2019 3:22 pm
I've know places 'sneak' these in a printer toner, stationary and petty cash - its better for IT to work with them rather than try to stop them by paperwork / board discussions.
Around here schools are so short on operating cash that they ask parents for donations of copy paper and toilet paper. I suppose donations of Raspberry Pi could also be solicited, but some parents may send the edible kind. It should be emphasized that the difficulty is not lack of money, only that allocating budgets for spending is highly bureaucratic.

Andyroo
Posts: 3306
Joined: Sat Jun 16, 2018 12:49 am
Location: Lincs U.K.

Re: is the pi worth it?

Thu May 16, 2019 3:38 pm

ejolson wrote:
Thu May 16, 2019 3:32 pm
Andyroo wrote:
Thu May 16, 2019 3:22 pm
I've know places 'sneak' these in a printer toner, stationary and petty cash - its better for IT to work with them rather than try to stop them by paperwork / board discussions.
Around here schools are so short on operating cash that they ask parents for donations of copy paper and toilet paper. I suppose donations of Raspberry Pi could also be solicited, but some parents may send the edible kind. It should be emphasized that the difficulty is not lack of money, only that allocating budgets for spending is highly bureaucratic.
(start of rant) Unfortunately I can believe that - shows how stupid humans are to limit the next generations funds :twisted: (end of rant)
Need Pi spray - these things are breeding in my house...

stuartiannaylor

Re: is the pi worth it?

Thu May 16, 2019 4:51 pm

ejolson wrote:
Thu May 16, 2019 2:59 pm

Using m.2 instead of USB3 sounds a bit like when IBM replaced the reverse-compatible AT bus with micro-channel. C
There is no reverse engineering with pcie, its native "PCI Express (PCIe) is the de-facto chip-to-chip connectivity standard for a wide range of applications from high-performance CPUs, networking, storage devices to battery-powered mobile devices. PCIe was first known as a board level bus system in personal computers, but today, with its wider links, distributed computing capabilities, and higher data rates, PCIe enables external connectivity in SoCs for high-performance servers"

m.2 is just 4x v2 pcie lanes just got a fancy name for a micro connector. Usb is just a hardware level generally on top of pci-e, but the quote above is from Arm not me.

User avatar
mahjongg
Forum Moderator
Forum Moderator
Posts: 11843
Joined: Sun Mar 11, 2012 12:19 am
Location: South Holland, The Netherlands

Re: is the pi worth it?

Thu May 16, 2019 5:39 pm

Andyroo wrote:
Thu May 16, 2019 3:38 pm
(start of rant) Unfortunately I can believe that - shows how stupid humans are to limit the next generations funds :twisted: (end of rant)
You know you are now so far off topic that I'm incline to lock the thread.... :mrgreen:

Andyroo
Posts: 3306
Joined: Sat Jun 16, 2018 12:49 am
Location: Lincs U.K.

Re: is the pi worth it?

Thu May 16, 2019 6:08 pm

mahjongg wrote:
Thu May 16, 2019 5:39 pm
Andyroo wrote:
Thu May 16, 2019 3:38 pm
(start of rant) Unfortunately I can believe that - shows how stupid humans are to limit the next generations funds :twisted: (end of rant)
You know you are now so far off topic that I'm incline to lock the thread.... :mrgreen:
Apologies - feel free to delete it. Wrists slapped deservedly. :oops:
Need Pi spray - these things are breeding in my house...

jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 22697
Joined: Sat Jul 30, 2011 7:41 pm

Re: is the pi worth it?

Thu May 16, 2019 6:25 pm

To answer the original question.

Yes.


Evidence. 26M sold, from the original (and quite slow) to the Pi3b+. A range of devices with obvious forward and backwards compatibility -the current Raspbian release works on those original Pi's, and HATS work on all the 40pin devices (Our one major not quite backward compatible step was moving from 26 to 40 pinsGPIO header).
Principal Software Engineer at Raspberry Pi (Trading) Ltd.
Contrary to popular belief, humorous signatures are allowed. Here's an example...
"My grief counseller just died, luckily, he was so good, I didn't care."

jdb
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 1999
Joined: Thu Jul 11, 2013 2:37 pm

Re: is the pi worth it?

Thu May 16, 2019 6:28 pm

stuartiannaylor wrote:
Thu May 16, 2019 4:51 pm
ejolson wrote:
Thu May 16, 2019 2:59 pm

Using m.2 instead of USB3 sounds a bit like when IBM replaced the reverse-compatible AT bus with micro-channel. C
There is no reverse engineering with pcie, its native "PCI Express (PCIe) is the de-facto chip-to-chip connectivity standard for a wide range of applications from high-performance CPUs, networking, storage devices to battery-powered mobile devices. PCIe was first known as a board level bus system in personal computers, but today, with its wider links, distributed computing capabilities, and higher data rates, PCIe enables external connectivity in SoCs for high-performance servers"

m.2 is just 4x v2 pcie lanes just got a fancy name for a micro connector. Usb is just a hardware level generally on top of pci-e, but the quote above is from Arm not me.
Your first paragraph does not address the point in the quoted text. The AT (ISA) bus was downgradeable, in that you could add PC 8-bit cards and they would fit in 16-bit busmastering slots and "just work", for some variable definition of "work".

The m.2 form factor has a connector that is explicitly an *internal* design, as in the socket is designed to only have things added/removed while the device is powered off. Every connector on the Pi is a hotpluggable connector - with the exception of GPIO, but with care these pins are hotpluggable with jumper wires (but not HATs). There are other issues with M.2 such as the Z-height of the connector, the requirement to drill a big fat hole for mounting, and the fact that the keep-out in the gap between the peripheral and the board also limits z-height of any components placed underneath. It is no panacea, and would be far down the list for inclusion on any theoretical Pi N+1 that had a PC-style expansion capability.

Ironically, the PCIe vertical socket was designed with hotplug in mind. The edge connector specifies the necessary extended ground/power fingers for sequencing - but I'd shy away from hotplugging a GPU board.
Rockets are loud.
https://astro-pi.org

stuartiannaylor

Re: is the pi worth it?

Thu May 16, 2019 7:21 pm

If the crux of the argument against m.2 is due to holes, I think I will rest my case.
From slight concern, I am beginning to get extremely worried in terms of some "engineers" responses.

The RockPro64 has a full size x4 PCIE vert its huge and useless as there is nothing to support card and board, great idea to have PCIE but ooof that doesn't work .
The RockPi4 has a singular m.2 but don't fancy trying the distro's on offer even if Pine from experience are not that bad, the actual Raxda layout is pretty cool though, but no solid distro like Raspbian.

Why "hot plugable" has become part of the discuss or even if its needed or anyone said get rid of the ethernet and USB.
No one said get rid of the GPIO or even said it had to be M.2, but just can not think of another that already has a plethora of cards available and connectors at relative low price.
The cameras, audio and micro SD card there is a load that could be thought of to provide the next generation educational SoC SBC in form factor.

If your focus of importance is continuation of commercial market, then thats your choice, but z-height, holes :) What sort of arguments are those.
It is off topic though so I guess here we can argue anything, but really :) lols.

Anyway anyone no why USB boot doesn't work when you swap the drive from a Pi3A+ to a Pi3B as with the price of SD cards and availability sod using them pesky fragile MicroSDs?

ejolson
Posts: 3039
Joined: Tue Mar 18, 2014 11:47 am

Re: is the pi worth it?

Thu May 16, 2019 7:25 pm

jdb wrote:
Thu May 16, 2019 6:28 pm
Every connector on the Pi is a hotpluggable connector - with the exception of GPIO, but with care these pins are hotpluggable with jumper wires (but not HATs).
I've heard of people damaging their Pi trying to hot plug the camera connector. For the record and for those who might otherwise try, is it possible to hot plug the camera connector on the current Pi or not?

Return to “Off topic discussion”