It seems the main area of improvement could be in the simplification of the development environments.
It also seems that silicon complexity is being hidden somewhat by GUI setup menus, which suggests that peripherals are not easily (re) configurable at runtime.
I recently put into production a USB Host project based around PIC24 series, what a PITA it was to set up, what with all the registers and so on.
You really can't beat the simplicity of the Propeller and the speed of development, it's so clean and simple.
People will always want to leverage existing experience before being forced to gain new experience. And people only want to gain new experience if they believe that it's beneficial.
This partly explains why people have problems with the IDEs. They are forced to work in a new environment instead of using the tools they are already familiar with. And, unless those IDEs are spectacularly better than what people are already using, they see little benefit in being forced to use those IDEs. Hence the negative opinions.
And that brings me to Spin. No one comes to the Propeller platform with existing Spin experience. Originally, in order to use the Propeller, you had to learn Spin. But you can't apply your Spin experience to other platforms and Spin itself doesn't provide a significantly better programming experience than other well-known languages do, meaning that there's not much benefit in learning Spin. Of course, this has been less of an issue now that C/C++ is ported, but the Propeller documentation clearly focuses on Spin. With the P2, it seems that supporting well known languages would be much more successful than providing an updated Spin. I understand that Spin still needs to be supported on the P2 (as a migration path for P1 users), but it should not be the focus. I think you will get a lot more attention if the P2 supported a a well-known language like Lua or JS.
You need added value. Pi is dead cheap. No money to make. But ADC, camera, i/o can be more expensive, so a propeller hat also
Got it. Yes.
@Seairth
Yes. Where spectacularly better IDE also includes spectacularly simple. Like the Adruino or Espruino systems. Dead easy to install. Almost nothing to configure, code and go.
Parallax started with small cool 'boards' for making things... Here we have a huge explosion of that with Pi and Arduino and where are the Prop products to leverage those? The Parallax products have stagnated while the world has jumped forward. Chips dreaming of a P2 for the last, what, 14 years(?), has left Parallax directionless and without creative vision. Just look at Adafruit and Sparkfun... that is what the Parallax should have been, you were there first!
We need practical Prop and P2 based add on boards for Pi, etc. As Erna said add on boards, that add unique capabilities, can sell for pretty good money. I just spent $50.00 for a motor board to pop on top of an Arduino Due to use for testing prototype custom CNC control software. Eventually that board will be replaced with a custom board with I/O expansion on it, done with cheap ready to use Arduino boards all communicating with serial connections. This project started just before the P2 Hot blew up and was supposed to be using a P2 + 2xP1... Now it is a Pi2 + 2x Arduino Due + Mega.
It coulda, shoulda, woulda been better with the Propellers...
As far as I can tell prices have been crashing towards zero. Chip makers have razor thin margins. Atmel had to throw in the towel and get absorbed my Microchip.
Higher up, Intel has just decided to cut it's work force by 12000.
Parallax has no hope of ever recouping the P2 development costs, at least not by selling chips. If they could sell them for 10 dollars and make a dollar on each that's, what was it?, 4 million chips to shift. I don't see that happening.
There needs to be clear and measurable value. And it looks like that will have to come from the board it sits on and the super slick software available for it that shows what it can do. Including ready made plug'n'play functionality that users of other systems, Arduino, Raspi, etc can make use of.
Maybe things are trending such that single chips aren't looked at as solutions as much as board-level products are, nowadays. Maybe the new people using electronics are more about hooking up wires than making PCB's from scratch.
Making PCBs seems to be more alive and well than ever, actually. The hobby electronics forums I visit (and lots of blogs and other pages) give me the impression that I'm now the odd one as I'm not designing PCBs and getting them produced by the myriad of very cheap good PCB houses. I see young people just starting out with some microprocessor or mcu, teenagers, designing PCBs before they're even fully aware that decoupling capacitors are supposed to be connected in parallel between VCC and ground, not in series with VCC..
It's never been easier or more convenient to create PCBs than now, and I wouldn't worry about that part at least.
The big change in the last 10 years has been true single chip operation. Only a single supply required and that compatible with Li batteries. No external boot memory, oscillator, or other devices. So basically the layout is connect up the lines, add a few bypass caps, USB, TTL serial or JTAG for in circuit programming and away you go. From a hardware standpoint the devices are simple to use. And as Brian mentioned above you can treat the cheap ones like logic gates and add multiple CPUs to the board simply.
The range in price has gone towards 0, with a low end 32bit CPU with 32K Flash and 8K RAM around 50 cents.
At the probable P2 price around $10 you now get a 65 nm design with a 300 MHz ARM, single and dual precision FPUs, cached 512K Flash, 256K RAM, USB, QSPI, SDRAM, LCD, camera, multiple UARTs, I2C, SPI, A/D and DAC. And it consumes just 90 mA running at full speed, with power down modes around 1uA.
On the tool side, yes Keil and Eclipse are frustrating, partly as they support all chips known to man so there are so many options. But even here there is progress with simpler mBed tools running on web which produce an image that is copied onto the "disk drive" that the device looks like to USB in download mode. This is being picked up by multiple vendors.
Are they fun? Not sure how to measure that, but I think they are relatively simple to use. Examples of peripheral programming are getting better, though they still can have a ways to go. For complex peripherals like Ethernet or USB, there are often ROM drivers built in for easier use.
I was thinking today that a Prop2 chip-resident editor, compiler, and debugger tool may take 32KB, which is about 1/30,000th of the size of the standard IDEs nowadays. I like the idea of getting right to work and not being bogged down by PC-level development software+OS issues. I wonder if that kind of thing would matter to many people....
That Espruino sounds pretty clean.
you don't have to look at another MCU
just see what can be done today with an interactive system named TACHYON-FORTH on a Propeller 1 and 2-FPGA.
Personally I would be lost if my development environment did not have web access. Either to look up what chips are available, or download their spec sheets.
A smart phone makes a great controller for devices these days, but a little small for reading chip PDFs, so I have that phone and a convertible tablet.
I'm not suggesting the IDE contain a web browser, just that whatever I am running my IDE on also has to have a web browser.
Many years ago we tried to do an IDE on the web, the problem is with even the fastest PC/Linux browsers are pretty bloated and slow to react. It is quite frustrating to try to text edit over the web. Now maybe you could do compiles over the web, as that would eliminate the user needing to setup the IDE, but frankly I still like local IDEs over the cloud. mBed does some of this if you want to see an example of a Cloud IDE.
I wonder how many people will really use an on-chip development environment seriously. It seems things are moving in the opposite direction. As Heater always points out, lots of tools are web-based with data in the cloud these days. They aren't even on a local PC much less on the target hardware itself. While I think it would be very cool to have self-hosted development capability, I find it hard to believe that it will be a strong selling point for P2.
Browsers on Linux are no more bloated or slow to react that those on Windows or Mac. Chrome is Chrome Firefox is Firefox, the operating systems are similarly efficient.
Editing text in a browser is quite OK, given a common browser and OS.
Never mind compiles over the net, we can compile Spin, for example in the browser itself. At speeds comparable to using a Spin compiler running as native code.
One does not need the "cloud" to use a web based IDE.
This is very cool! I had forgotten about this work. Is there a github repository that contains the current code? How hard is it to use Emscripten to convert a C program to Javascript?
Even the interpreted environments that are being pushed to micros (espruino, micro-python, elua (nodemcu), MicroMite BASIC, etc.) are coupled with some sort of IDE more and more. While you can directly interface with the interpreter via a serial connection if you want/need to, the IDE provides support for all the things you want in an IDE (files, code repositories, full screen, code completion, etc.). If a browser based IDE can become the "new" Eclipse/Arduino for these little guys, I won't complain.
IIDE - Interactive Integrated Development Environments!
As a hobbyist, there is no other microcontroller that can be directly compared with the P1. So, I think it is fair to say that the P2 is going to be special.
What Heater said: "A decade or so ago electronics as a hobby was almost dead."
At that time, I was trying to find a way into the hobby. I did a lot of reading, but when it came time to actually pulling the trigger, I couldn't. It didn't make any sense: all of the things that I wanted to do seemed impossible, and what was possible seemed so limited that I could not imagine it holding my interest. The hobby wasn't dead, but for me it was practically unapproachable.
I literally stumbled across Parallax's Forum.
Spin is so simple, I didn't need to read anything. The Forum was so good that most of my questions had already been asked and answered repeatedly. I didn't have to go anywhere else. What more could a hobbyist ask for? For me it was all about sensors. All I had to do was buy one, fiddle with the code a bit and it became clear as crystal. Quite an amazing experience.
If the Arduino had been around at that time, I would have considered it... and I might have made a mistake. This is still an issue. People finding the Arduino and being confused by it. A migration path for these people needs to be made as easy as possible. I wouldn't lose any sleep over it, but it is there as an issue.
Education and engineering are different kettles of fish. They have different needs, which are well represented here.
Raspberry Pi's: a computer on a card. Sounded perfect for projects. Required a different operating system, with its own care and feeding. Mine are collecting dust. Now Pi's have iWin10... but so do other cheap computers with smaller form factors, which are more powerful out of the box and are a lot easier to use. A P2 and a cheap Win10 box seem like an unbeatable development platform... (except it is hard to put a hat on a Win10 box... but why would you want to?)
Tools
As an amateur, the only tool I really need is a scriptable text editor with sliders that allow me to scroll my PASM2 code.
In the future, I would like to be able to jump around PASM2 code so that I don't get blisters on my middle finger. It would be nice if the editor was smart enough to show me all of the lines of code that use a particular variable, have a particular comment or keyword. It would be nice to have routines listed and reachable by clicking.
Sensors
The sensors that surround the P2 will determine the level of success of the P2.
Exciting new sensors seem to be popping out of the woodwork. As a hobbyist, I have to have them, BUT the really good ones all come with... API's. Can you imagine it?
API's... why in the world would I want to use an API?
What kind of tool chain do I need for THAT? (I know but I'm not happy with the answer)
If I want to understand what a sensor is doing... I don't need an API, I need Parallax.
.
I'd like to add a +1 to the Pi Hat idea. That is something Parallax can get in production right now with the P1, then transition the design to the P2.
.
For it to be successful, I think it needs to have the following attributes:
1) A full Python library, installable via "pip" or other means onto the Raspberry Pi to control the Prop. With a Python library, Parallax would be leveraging a popular language that people already have an IDE and are familiar. It may be time to let Spin go and concentrate on languages folks are already familiar.
2) As many "real world" interfaces as you can fit on it - ADC, motor controller, servo output pins, speaker, WS-LED, a few 5V level shifted GPIO pins, ect. Maybe a compass. Maybe Bluetooth and a smartphone app the Prop can talk to (I think the Pi3 already has Bluetooth, though).
3) A DC converter to power the Rpi from the Propeller hat, to simplify power hookups.
4) Optical isolation between the Prop and the Rpi. Use the DIP version of the Prop so if the user blows up the Prop, the Rpi is protected and the Prop is easily replaceable.
5) Start students using Python to control the Prop, then slowly transition them to running C++ directly on the Prop. Avoid Spin, it will only complicate matters and their teachers won't take the time to learn it.
6) Many step by step tutorials and projects utilizing the new "hat" - emphasizing the Rpi as the "thinker" and the Prop as the "doer". One of Adafruits big strengths (IMHO) is that on the same page as every product is a LEARN link to a blog post showing step by step instruction how to use the product. Their engineers make videos showing live how to hook up the products and get the results the purchaser wants.
It also seems that silicon complexity is being hidden somewhat by GUI setup menus, which suggests that peripherals are not easily (re) configurable at runtime.
Our customers developing smartphones couldn't even change baud rates on the fly. A lot of designs are based on reference platforms these days, and there is a lot of plug and pray engineering.
Also silicon providers supply application engineers to do a lot the design work for customers.
One funny thing about the 1000 page documents - we once had a customer print out chip documentation from us and a competitor, and show that the competitor's was better because it weighed more. (This was in Japan.)
.
I'd like to add a +1 to the Pi Hat idea. That is something Parallax can get in production right now with the P1, then transition the design to the P2.
.
For it to be successful, I think it needs to have the following attributes:
1) A full Python library, installable via "pip" or other means onto the Raspberry Pi to control the Prop. With a Python library, Parallax would be leveraging a popular language that people already have an IDE and are familiar. It may be time to let Spin go and concentrate on languages folks are already familiar.
2) As many "real world" interfaces as you can fit on it - ADC, motor controller, servo output pins, speaker, WS-LED, a few 5V level shifted GPIO pins, ect. Maybe a compass. Maybe Bluetooth and a smartphone app the Prop can talk to (I think the Pi3 already has Bluetooth, though).
3) A DC converter to power the Rpi from the Propeller hat, to simplify power hookups.
4) Optical isolation between the Prop and the Rpi. Use the DIP version of the Prop so if the user blows up the Prop, the Rpi is protected and the Prop is easily replaceable.
5) Start students using Python to control the Prop, then slowly transition them to running C++ directly on the Prop. Avoid Spin, it will only complicate matters and their teachers won't take the time to learn it.
6) Many step by step tutorials and projects utilizing the new "hat" - emphasizing the Rpi as the "thinker" and the Prop as the "doer". One of Adafruits big strengths (IMHO) is that on the same page as every product is a LEARN link to a blog post showing step by step instruction how to use the product. Their engineers make videos showing live how to hook up the products and get the results the purchaser wants.
Don't just throw out a "hat" and a datasheet.
My $0.02 . . .
This sounds awesome, but more appropriate for the P1 to me. The p2 is going to be so immensely powerful, that I wouldn't want to use it in such a basic learning environment that you describe. With 512kB of code space, you can run the Python interpreter directly on the chip - no need for Python on the Pi. However, writing a Python library for the Pi that communicates with a P1 to control servos/LEDs/etc sounds much more useful.
Sadly, I don't think either one of them have benefitted from the 10 million Raspberry Pis sold in the numbers they may have expected.
It's more market penetration and name recognition than anything. Lots of people have heard about Arduino....propellers are on boats and airplanes. Now it's all about the ESP8266 and relatives - ok, a $5 (now $2.30) WiFi enabled micro has a pretty good hook to get people (hobbyists and professionals) interested and willing to try it for the prices of a burger.
The crowds aren't pounding at the doors demanding the P2 get released...or even the P1+...or even the trusty old P1.
Where is the P2 being pitched? At the hobbyist market or commercial market?
They are two very different markets! Traditionally Parallax's market was the hobbyist.
I have been a commercial designer for more than 40 years, although mostly these days I am just a hobbyist. I can see commercial markets for hobbyists for a P1+ right now. I know of commercial embedded designs for P1+ right now.
Will there be a commercial market for hobbyists or commercial embedded designs in 2017+ when P2 silicon is available, or much later when the tools to support P2 are done ???
The market has moved on since P1 & P2 were conceived. P1 needs a refresh (P1+) now before it is too late! P2 is not a refresh. It may as well be considered a new design as its too far away from P1 now.
Just the other day I picked up my very old pre P2Hot debugger code. I wrote it with backward porting to P1 in mind so it mostly used P1 style instructions.
I went to convert it to the new P2 instruction set. What a nightmare it turned out to be. No more waitcnt count,delay. It is now addctn X,delay followed by waitctn. Calls are completely different, as are ret. there were others too. I had to continually refer to the online write up.
Of course these things will become second nature soon enough. I am just pointing out its the basics that have changed and they might scare existing users who want to convert to P2. For them a P1+ would make more sense at this time, and after all it's those current commercial users who will be the initial volume users of a new chip.
As my marketing consultant used to say, "it is six times harder to get a new customer than it is to keep an existing customer".
About 5 years ago, Ken asked if they should do the P1B or wait a little longer for the imminent P2. Unfortunately for us, we all (me included) said wait for P2. In hindsight, what a mistake that was! P1B would have overcome the I/O problem, so we could have external SRAM. I am certain had the P1B been done, P1+ would have also been here long ago.
There are ever more choices for micropower operation. I was yesterday evaluating the PIC16F18313 8-pin for a key-size logger I've been asked to build. It does 250nA with the watchdog running. Operation at 8µA at 32kHz, which is similar to what I get with the Prop I on RCslow. @Chip, from what I gather the new Prop will not have micropower modes, but I'm not sure. I've been following the saga loosely.
For me these decisions are based largely on the learning curve and a strategy to avoid the risk of misunderstanding the data sheets or the hidden bugs and gotchas. I pick one or two and stick with them. Gosh golly the data sheet for the PIC16F18313 weighs in at 442 pages, far more than the PIC12F675 that I know well. It even has "peripheral pin remapping", but that is surrounded by ifs-ands-buts, hey! recall the total flexibility of Prop (old and new) pins. The number and complexity of modules is mind-boggling, well documented, intriguing, a possible adventure with risks and rewards. Should the old dog learn new tricks? Well, learning microcontrollers isn't really my thing. I just want to do things with them. A dilemma.
This topic is making me nervous. It almost makes me think that Chip is having second thoughts about the P2 design. Are we still good to go with P2 as it is? Or is another redesign being contemplated?
Comments
Without customers who will buy the P2 in 10k/pa quantities then there's little point in it ever being made into Silicon.
I recently put into production a USB Host project based around PIC24 series, what a PITA it was to set up, what with all the registers and so on.
You really can't beat the simplicity of the Propeller and the speed of development, it's so clean and simple.
Hopefully P2 can carry on in the same way.
This partly explains why people have problems with the IDEs. They are forced to work in a new environment instead of using the tools they are already familiar with. And, unless those IDEs are spectacularly better than what people are already using, they see little benefit in being forced to use those IDEs. Hence the negative opinions.
And that brings me to Spin. No one comes to the Propeller platform with existing Spin experience. Originally, in order to use the Propeller, you had to learn Spin. But you can't apply your Spin experience to other platforms and Spin itself doesn't provide a significantly better programming experience than other well-known languages do, meaning that there's not much benefit in learning Spin. Of course, this has been less of an issue now that C/C++ is ported, but the Propeller documentation clearly focuses on Spin. With the P2, it seems that supporting well known languages would be much more successful than providing an updated Spin. I understand that Spin still needs to be supported on the P2 (as a migration path for P1 users), but it should not be the focus. I think you will get a lot more attention if the P2 supported a a well-known language like Lua or JS.
@Seairth
Yes. Where spectacularly better IDE also includes spectacularly simple. Like the Adruino or Espruino systems. Dead easy to install. Almost nothing to configure, code and go.
I think you are right about Spin.
Parallax started with small cool 'boards' for making things... Here we have a huge explosion of that with Pi and Arduino and where are the Prop products to leverage those? The Parallax products have stagnated while the world has jumped forward. Chips dreaming of a P2 for the last, what, 14 years(?), has left Parallax directionless and without creative vision. Just look at Adafruit and Sparkfun... that is what the Parallax should have been, you were there first!
We need practical Prop and P2 based add on boards for Pi, etc. As Erna said add on boards, that add unique capabilities, can sell for pretty good money. I just spent $50.00 for a motor board to pop on top of an Arduino Due to use for testing prototype custom CNC control software. Eventually that board will be replaced with a custom board with I/O expansion on it, done with cheap ready to use Arduino boards all communicating with serial connections. This project started just before the P2 Hot blew up and was supposed to be using a P2 + 2xP1... Now it is a Pi2 + 2x Arduino Due + Mega.
It coulda, shoulda, woulda been better with the Propellers...
As far as I can tell prices have been crashing towards zero. Chip makers have razor thin margins. Atmel had to throw in the towel and get absorbed my Microchip.
Higher up, Intel has just decided to cut it's work force by 12000.
Parallax has no hope of ever recouping the P2 development costs, at least not by selling chips. If they could sell them for 10 dollars and make a dollar on each that's, what was it?, 4 million chips to shift. I don't see that happening.
There needs to be clear and measurable value. And it looks like that will have to come from the board it sits on and the super slick software available for it that shows what it can do. Including ready made plug'n'play functionality that users of other systems, Arduino, Raspi, etc can make use of.
It's never been easier or more convenient to create PCBs than now, and I wouldn't worry about that part at least.
The range in price has gone towards 0, with a low end 32bit CPU with 32K Flash and 8K RAM around 50 cents.
At the probable P2 price around $10 you now get a 65 nm design with a 300 MHz ARM, single and dual precision FPUs, cached 512K Flash, 256K RAM, USB, QSPI, SDRAM, LCD, camera, multiple UARTs, I2C, SPI, A/D and DAC. And it consumes just 90 mA running at full speed, with power down modes around 1uA.
On the tool side, yes Keil and Eclipse are frustrating, partly as they support all chips known to man so there are so many options. But even here there is progress with simpler mBed tools running on web which produce an image that is copied onto the "disk drive" that the device looks like to USB in download mode. This is being picked up by multiple vendors.
Are they fun? Not sure how to measure that, but I think they are relatively simple to use. Examples of peripheral programming are getting better, though they still can have a ways to go. For complex peripherals like Ethernet or USB, there are often ROM drivers built in for easier use.
just see what can be done today with an interactive system named TACHYON-FORTH on a Propeller 1 and 2-FPGA.
A smart phone makes a great controller for devices these days, but a little small for reading chip PDFs, so I have that phone and a convertible tablet.
Vim open in one window, Chrome open in another.
No need to build that into any IDE.
Except, why not build the IDE into the browser?
Many years ago we tried to do an IDE on the web, the problem is with even the fastest PC/Linux browsers are pretty bloated and slow to react. It is quite frustrating to try to text edit over the web. Now maybe you could do compiles over the web, as that would eliminate the user needing to setup the IDE, but frankly I still like local IDEs over the cloud. mBed does some of this if you want to see an example of a Cloud IDE.
Wait a minute...
Browsers on Linux are no more bloated or slow to react that those on Windows or Mac. Chrome is Chrome Firefox is Firefox, the operating systems are similarly efficient.
Editing text in a browser is quite OK, given a common browser and OS.
Never mind compiles over the net, we can compile Spin, for example in the browser itself. At speeds comparable to using a Spin compiler running as native code.
One does not need the "cloud" to use a web based IDE.
See our proof of concept, experimental Spin IDE here: http://propanel.oo.fi/lab/spine.html
OK I'm impressed, JS has gotten better. And yes you are really running this locally though served by a webpage.
But do you really expect this to run on a P2?
IIDE - Interactive Integrated Development Environments!
What Heater said: "A decade or so ago electronics as a hobby was almost dead."
At that time, I was trying to find a way into the hobby. I did a lot of reading, but when it came time to actually pulling the trigger, I couldn't. It didn't make any sense: all of the things that I wanted to do seemed impossible, and what was possible seemed so limited that I could not imagine it holding my interest. The hobby wasn't dead, but for me it was practically unapproachable.
I literally stumbled across Parallax's Forum.
Spin is so simple, I didn't need to read anything. The Forum was so good that most of my questions had already been asked and answered repeatedly. I didn't have to go anywhere else. What more could a hobbyist ask for? For me it was all about sensors. All I had to do was buy one, fiddle with the code a bit and it became clear as crystal. Quite an amazing experience.
If the Arduino had been around at that time, I would have considered it... and I might have made a mistake. This is still an issue. People finding the Arduino and being confused by it. A migration path for these people needs to be made as easy as possible. I wouldn't lose any sleep over it, but it is there as an issue.
Education and engineering are different kettles of fish. They have different needs, which are well represented here.
Raspberry Pi's: a computer on a card. Sounded perfect for projects. Required a different operating system, with its own care and feeding. Mine are collecting dust. Now Pi's have iWin10... but so do other cheap computers with smaller form factors, which are more powerful out of the box and are a lot easier to use. A P2 and a cheap Win10 box seem like an unbeatable development platform... (except it is hard to put a hat on a Win10 box... but why would you want to?)
Tools
As an amateur, the only tool I really need is a scriptable text editor with sliders that allow me to scroll my PASM2 code.
In the future, I would like to be able to jump around PASM2 code so that I don't get blisters on my middle finger. It would be nice if the editor was smart enough to show me all of the lines of code that use a particular variable, have a particular comment or keyword. It would be nice to have routines listed and reachable by clicking.
Sensors
The sensors that surround the P2 will determine the level of success of the P2.
Exciting new sensors seem to be popping out of the woodwork. As a hobbyist, I have to have them, BUT the really good ones all come with... API's. Can you imagine it?
API's... why in the world would I want to use an API?
What kind of tool chain do I need for THAT? (I know but I'm not happy with the answer)
If I want to understand what a sensor is doing... I don't need an API, I need Parallax.
I'd like to add a +1 to the Pi Hat idea. That is something Parallax can get in production right now with the P1, then transition the design to the P2.
.
For it to be successful, I think it needs to have the following attributes:
1) A full Python library, installable via "pip" or other means onto the Raspberry Pi to control the Prop. With a Python library, Parallax would be leveraging a popular language that people already have an IDE and are familiar. It may be time to let Spin go and concentrate on languages folks are already familiar.
2) As many "real world" interfaces as you can fit on it - ADC, motor controller, servo output pins, speaker, WS-LED, a few 5V level shifted GPIO pins, ect. Maybe a compass. Maybe Bluetooth and a smartphone app the Prop can talk to (I think the Pi3 already has Bluetooth, though).
3) A DC converter to power the Rpi from the Propeller hat, to simplify power hookups.
4) Optical isolation between the Prop and the Rpi. Use the DIP version of the Prop so if the user blows up the Prop, the Rpi is protected and the Prop is easily replaceable.
5) Start students using Python to control the Prop, then slowly transition them to running C++ directly on the Prop. Avoid Spin, it will only complicate matters and their teachers won't take the time to learn it.
6) Many step by step tutorials and projects utilizing the new "hat" - emphasizing the Rpi as the "thinker" and the Prop as the "doer". One of Adafruits big strengths (IMHO) is that on the same page as every product is a LEARN link to a blog post showing step by step instruction how to use the product. Their engineers make videos showing live how to hook up the products and get the results the purchaser wants.
Don't just throw out a "hat" and a datasheet.
My $0.02 . . .
Our customers developing smartphones couldn't even change baud rates on the fly. A lot of designs are based on reference platforms these days, and there is a lot of plug and pray engineering.
Also silicon providers supply application engineers to do a lot the design work for customers.
One funny thing about the 1000 page documents - we once had a customer print out chip documentation from us and a competitor, and show that the competitor's was better because it weighed more. (This was in Japan.)
This sounds awesome, but more appropriate for the P1 to me. The p2 is going to be so immensely powerful, that I wouldn't want to use it in such a basic learning environment that you describe. With 512kB of code space, you can run the Python interpreter directly on the chip - no need for Python on the Pi. However, writing a Python library for the Pi that communicates with a P1 to control servos/LEDs/etc sounds much more useful.
The trouble is, that market is now full of cheap non-AVR Arduino form factor boards based around everything from the Cortex M0 all the way up.
Pimoroni Propeller Hat available through Parallax or Pimoroni.
Bill Henning's RoboPi Robot Controller for the Raspberry Pi availble from http://www.mikronauts.com/
Sadly, I don't think either one of them have benefitted from the 10 million Raspberry Pis sold in the numbers they may have expected.
It's more market penetration and name recognition than anything. Lots of people have heard about Arduino....propellers are on boats and airplanes. Now it's all about the ESP8266 and relatives - ok, a $5 (now $2.30) WiFi enabled micro has a pretty good hook to get people (hobbyists and professionals) interested and willing to try it for the prices of a burger.
The crowds aren't pounding at the doors demanding the P2 get released...or even the P1+...or even the trusty old P1.
Where is the P2 being pitched? At the hobbyist market or commercial market?
They are two very different markets! Traditionally Parallax's market was the hobbyist.
I have been a commercial designer for more than 40 years, although mostly these days I am just a hobbyist. I can see commercial markets for hobbyists for a P1+ right now. I know of commercial embedded designs for P1+ right now.
Will there be a commercial market for hobbyists or commercial embedded designs in 2017+ when P2 silicon is available, or much later when the tools to support P2 are done ???
The market has moved on since P1 & P2 were conceived. P1 needs a refresh (P1+) now before it is too late! P2 is not a refresh. It may as well be considered a new design as its too far away from P1 now.
Just the other day I picked up my very old pre P2Hot debugger code. I wrote it with backward porting to P1 in mind so it mostly used P1 style instructions.
I went to convert it to the new P2 instruction set. What a nightmare it turned out to be. No more waitcnt count,delay. It is now addctn X,delay followed by waitctn. Calls are completely different, as are ret. there were others too. I had to continually refer to the online write up.
Of course these things will become second nature soon enough. I am just pointing out its the basics that have changed and they might scare existing users who want to convert to P2. For them a P1+ would make more sense at this time, and after all it's those current commercial users who will be the initial volume users of a new chip.
As my marketing consultant used to say, "it is six times harder to get a new customer than it is to keep an existing customer".
About 5 years ago, Ken asked if they should do the P1B or wait a little longer for the imminent P2. Unfortunately for us, we all (me included) said wait for P2. In hindsight, what a mistake that was! P1B would have overcome the I/O problem, so we could have external SRAM. I am certain had the P1B been done, P1+ would have also been here long ago.
For me these decisions are based largely on the learning curve and a strategy to avoid the risk of misunderstanding the data sheets or the hidden bugs and gotchas. I pick one or two and stick with them. Gosh golly the data sheet for the PIC16F18313 weighs in at 442 pages, far more than the PIC12F675 that I know well. It even has "peripheral pin remapping", but that is surrounded by ifs-ands-buts, hey! recall the total flexibility of Prop (old and new) pins. The number and complexity of modules is mind-boggling, well documented, intriguing, a possible adventure with risks and rewards. Should the old dog learn new tricks? Well, learning microcontrollers isn't really my thing. I just want to do things with them. A dilemma.