Dave, 96 I/O pins plus power etc. of standard connectors is going to take quite a bit of board space anyway. The P1 QuickStart couldn't be made much smaller and still have its simple standard 0.1 inch pitch header. P2 will need a more elaborate power supply because 1.8V core. Putting a few cheap I/O components that can be ignored if not wanted makes the board dual-purpose in a very good way and really not that much bigger or expensive.
Personally, I'm more interested in the little module they talked about with just the P2, SDRAM and I forget what else.
These are a major pain for me to solder and so I'd very much like to use these little modules....
Dave, 96 I/O pins plus power etc. of standard connectors is going to take quite a bit of board space anyway.
Yes of course you are so right.
96 i/o plus..... how many total?
At 0.1 inches represents about ten inches of pins.
2 times 50; 4 times 25 or about 2.8 inches square (round the edge), is HUGE!!!!
Point taken:thumb:
Yes of course you are so right.
96 i/o plus..... how many total?
At 0.1 inches represents about ten inches of pins.
2 times 50; 4 times 25 or about 2.8 inches square (round the edge), is HUGE!!!!
Point taken:thumb:
Dave
Yes but if you're driving a TV or monitor it makes perfect sense to have something that screws into the Vesa mounts of it, which are either on a 75x75 or 100x100 mm square pitch.
We have to wait for accurate power consumption figures, but depending on cogs required and clock rate it may even be possible to draw 2.5 watts out of the USB socket that many tvs & monitors have onboard
One industry sector in which the P2 would be a natural fit is gaming equipment: video poker, slot machines, etc. These require not-too--hi-res video, decent sound effects. button and LED I/O, slip reading, slip printing, currency recognition, and networking capabilities. I think the P2 could be a one-chip solution to all those needs.
One industry sector in which the P2 would be a natural fit is gaming equipment: video poker, slot machines, etc. These require not-to--hi-res video, decent sound effects. button and LED I/O, slip reading, slip printing, currency recognition, and networking capabilities. I think the P2 could be a one-chip solution to all those needs.
-Phil
Would appreciate hearing more ideas like this one. Contrary to what Chip suggests, we will have some customer-focused presentations to sell the P2. No boring PPT, but engineering-ready appnotes and just the right amount of presentation slides for the managerial types. I would enjoy hearing all of your suggestions for the best applications of the P2. Please do share.
If you've been through a casino recently you might notice that "slot machines" are quickly morphing into generic video games, with high resolution dynamic graphics. When slot machines started to be computerized in the 1980's the regulators would conduct spot checks to pop the EPROMs and put them in a comparator box to make sure they hadn't been hacked. I'm not sure what they're doing nowadays but I'm sure it can't be the same, unless there is some kind of Chinese wall between the display module probably running Linux and the gaming module still running old tech. Anyway the regulatory environment in Nevada is key to entering the industry, since Nevada requires manufacturers whose products are approved for Nevada to comply with those same requirements globally whether other venues are as strict or not.
I was going to say. There are arcades of gambling machines all around here. You will find poker machines near the checkouts in supermarkets. All owned and run my a state run monopoly. Those machines can play a variety of games, newer ones have two big screens. For sure they are networked. If they are not running Windows or Linux underneatth I would be surprised.
In bars you will sometimes find games consoles on some of the tables. Again high res screens and networked for multi-players modes.
Just the other day I found one displaying some Windows "Fatal error" dialog box in the middle of the screen. I clicked the OK button and it continued to try and persuade me to play Trivial Pursuit by means of a nice looking girl in a video suggestively winking at me. Very anoying. I wished I'd left the fatal error dialog up.
The Props role in a gaming rig now a days might be in some sort of controller input device.
Just the other day I found one displaying some Windows "Fatal error" dialog box in the middle of the screen.
LOL =D
it reminds me a huge LED screen in Montreal where ads were displayed. One day, a big "Windows update" and restart notice popped up in the screen. shi* happens when you use Windows OS
There are arcades of gambling machines all around here ... All owned and run my a state run monopoly ... For sure they are networked. If they are not running Windows or Linux underneatth I would be surprised.
Slot machines are most definitely networked at the level of the player's club card reader and display, but if they're made in Nevada the gaming modules are most definitely not allowed to communicate with each other in any way. This doesn't mean someone outside Nevada isn't doing that but it would be considered sleazy in traditional industry circles.
The bar machines which aren't "gambling" devices because you can't win money from them, most definitely are networked, phone home periodically, and are running LInux. I saw an article about the biggest distributor some years ago and they were quite bullish on the ability to distribute new games and monitor income without physical site visits.
I just don't know how the compliance enforcement would work on a machine with a surface mounted flash chip that can be field reprogrammed but not mounted in a comparator to see if that's been done. Unfortunately with the decline of the industry and opportunities for advantage play I lost interest in keeping track about the time the technology started to change. But if a company like Parallax wants to make an entry into the gaming world, someone would have to find out what the new rules are.
As I live down the street from Nevada (literally 2 miles away), even the the cheap casinos have networked game machines, that is how they offer club cards, measure your activity and the like.
The big boys in gaming are IGT and Bally. IGT has a big operation in Reno. Last time I checked (about 5 years back) most machines had the gaming front end and a linux backend. The Linux part did all the communication, player tracking. Not sure where the P2 would fit into that, while yes it does video on the scale of retro games, can it do MP3 video decoding? Yes the P2 can do the older bar video poker games, that technology is well over 10 years old. Yes you could build a better version of that, but people designing those today are going to be using the technology available in any $100 Android tablet.
Parallax has to focus on what it has been good at, and use that. That has been the hobbyist education business. Is there some niche for the P2 out there, maybe. But the microprocessor field has always been very crowded. Look at the list of architectures that no longer exist or are so minimal in the market (Z80, 6800, 68000, 960, COPs, Alpha, 6500, 8051 and more that I've forgotten). The general market while large is coalescing around a couple architectures, ARM and maybe PICs at the real low end, though in my opinion the cost of the packaging will exceed the cost of the die and at that point the PICs will go away. We are probably less than a couple years from that with ARMs in volume well less than $1.
We are probably less than a couple years from that with ARMs in volume well less than $1.
Chips from the LPC800 family (Cortex M0+) are already in the $0.59 range, small qty. Really just depends on which ARM you want. There certainly is a broad spectrum of performance and cost. The NXP chip I use most frequently is $1.26, single piece. I haven't even thought of using a PIC in the last 10 years! Crazy how things evolve.
Look at the list of architectures that no longer exist or are so minimal in the market (Z80, 6800, 68000, 960, COPs, Alpha, 6500, 8051 and more that I've forgotten).
8051s are still alive and kicking. There are a dozen or so companies making them in one form or another. For one example, look at how well Silicon Labs have done in recent years, with 8051s the mainstay of their MCU product line.
Anyway, I agree that Parallax should continue to market products to the educator and hobbyist communities. But I do also think there are certain niche markets to be had in, eg, scientific and industrial applications.
Z80 and 6502 are still in current use too. In a lot of cases the cores are licensed and fabbed onto chips with other hardware like LCD controllers. It's easy to get fixated on things like PC's, smartphones, and game consoles that we recognize as "computers" and forget that every microwave oven, washing machine, and key ring photo album has a computer in it too -- and one that doesn't need to be very powerful.
Making that determination is actually a task with a complexity on par with developing the product! I've been through a couple of those exercises. It was very interesting! A lot of the problem boils down to what others see as value and coming to understand what that value is worth to them. You can get a baseline that way, and it's messy. Gotta talk to a lot of people, get data, run scenarios, produce reference designs, concepts, etc... A lot more of the problem comes down to understanding what the competing offerings are, which will indicate a variance on the price to compete. Finally, there is the risk costs. Product life window, time to market, etc... Hit early, price can be high, margins high, because early adopters are buying time and taking risks for their own reasons. Hit in the middle, price is solid, margins reasonable, etc... Late? Well, price is going to be lower, margins thinner.
Once all that is sorted out, price is somewhat known, and if that price works, do it. If not, value must be added so that price does work.
Wash, rinse, repeat over the life cycle of whatever it is.
Parallax has to focus on what it has been good at, and use that. That has been the hobbyist education business.
Hey Bruce! The group of "hobby education" is two different audiences with 20-30% overlap in our experience. Educational markets are addressed in person, by training teachers.
You omitted the commercial market as a focus of ours. Did you know that we have production customers using Propeller 1 who will also design in Propeller 2? These customers are an absolute necessity to our future success and will be the continued focus of our FAEs. The latest production effort is www.swapbox.com (shared with their approval, of course). Production customers who choose use the Propeller for rapid prototype to production, flexible I/O from software, ease of multicore, object support, personal FAE assistance, and it makes their products far more practical to develop. Their applications are in renewable energy, robotics, medical, environmental measurements and machine control. We needn't sell ARM volumes to this audience to prosper, only a few million chips a year once P2 is underway.
I think I understand your point, and simply want to point out that we have relevant production customers.
Sure you can find those dying MCUs out there, but wander around Design West (ESC), and is that what you see being offered. Other than Microchip with various PICs and a small remnant of Zilog (wow still with the Captain Zilog ad campaign)
But the question you have to answer, if you are starting a clean sheet design, would you really look at much other than an ARM for an embedded project? Sure the user manual is close to 1000 pages for many of them, but the difference between a hobby/education project and a project that will go into even limited production, is that a team of people working full time for a couple months gets amortized over the production run. The hobby project gets amortized over a couple units.
The ARMs will have a lower cost and are a 1 chip solution, no boot Flash or core power supply needed. While the prop can do limited specialized IO, it can't compete with dedicated hardware. Take a $8 ARM these days and compare to a P2. That ARM will do 4 UARTs, 100 Mb Ethernet, 2 CAN, 2 USBs, LCD/VGA controllers all built in and all capable of running in parallel. Now the P2 can do any couple of those, but not all simultaneously, and it still requires extra support hardware.
This is not to say there is no place for the P2, it's just unlikely to be a volume product. Maybe in the lab or education it will shine and capture enough of these to build a business. Maybe there is a market as a peripheral to be licensed to bigger vendors, though there are already homegrown versions of this at both NXP and TI.
Z80 and 6502 are still in current use too. In a lot of cases the cores are licensed and fabbed onto chips with other hardware like LCD controllers. It's easy to get fixated on things like PC's, smartphones, and game consoles that we recognize as "computers" and forget that every microwave oven, washing machine, and key ring photo album has a computer in it too -- and one that doesn't need to be very powerful.
Precisely. And it's nigh impossible to get pricing below that of an 8051 in such applications. They'll be around for a while.
You omitted the commercial market as a focus of ours.... These customers are an absolute necessity to our future success and will be the continued focus of our FAEs.
Ken: does Parallax have any independant sales representatives? I ask because, while seeming old-school, it's still a great way to get your ICs designed in. Those guys and gals are in places daily that you'd want a shot at doing business with, and often they have intimate knowledge of new designs in their early stages -- the very best time to sell the advantages of your particular solution. The only cost to you, generally, is a small commission *if* they get your parts designed in.
Perhaps you do already have sales reps out there, but when I checked the website(s) I couldn't find any mention of them.
Ken: does Parallax have any independant sales representatives? I ask because, while seeming old-school, it's still a great way to get your ICs designed in. Those guys and gals are in places daily that you'd want a shot at doing business with, and often they have intimate knowledge of new designs in their early stages -- the very best time to sell the advantages of your particular solution. The only cost to you, generally, is a small commission *if* they get your parts designed in.
Perhaps you do already have sales reps out there, but when I checked the website(s) I couldn't find any mention of them.
KC_Rob, no, we don't use sales representatives at the moment. I'm familiar with them from our own design-in processes when we source parts from Future, Arrow or other companies that use representatives. We are in talks with a distributor of this type who use representatives throughout the country, as this kind of approach would be more manageable than us working directly with individual representatives. We have only 42 people at Parallax so a representative arrangement would best be handled by a stocking distributor. I appreciate this thought and recognize the value of personal technical sales support.
KC_Rob, no, we don't use sales representatives at the moment. I'm familiar with them from our own design-in processes when we source parts from Future, Arrow or other companies that use representatives. We are in talks with a distributor of this type who use representatives throughout the country, as this kind of approach would be more manageable than us working directly with individual representatives. We have only 42 people at Parallax so a representative arrangement would best be handled by a stocking distributor. I appreciate this thought and recognize the value of personal technical sales support.
Gotcha. I was pretty certain that you had at least looked into this, and you have.
I will say, though, that I have worked for smaller companies (less than $10 million annually) that have managed to retain and work with sales reps throughout the US on their own without too much difficulty. Once the reps are selected and contracts in place, the process becomes somewhat routine.
@Brucee, the business is already running and viable on education. Parallax more or less set the bar on this early on. Education has specific requirements that pay well when met and that's a very significant part of what Parallax does. It just so happens that hobby, small entrepreneurial efforts dovetail nicely with that, all of which funded the P2 nicely enough. They know that market cold and are doing the work to realize success. This is a known to them, viable too.
Those efforts will be replicated on the P2, and in fact are happening right now. The years spent serving educators positions Parallax very well to compete in education / labs with P2. I would not characterize this as a "maybe", but more like "how much?" two entirely different things! Frankly, I think some of Chip's vision for the "bench computer" where visualizing real world phenomena has some serious appeal in education and labs. Some of the very best learning happens when we can get our senses wrapped around stuff. Being able to visualize, manipulate, capture data, produce feedback, interact, control is a very high value activity! We know it's high value because just look at the cost of gear designed to do just that. Look at what people will pay for visualization, just as one related example. In that world, people take large data and they break it down to a picture. The PhD people, can look at the data and visualize internally, sometimes... most everybody else can't most of the time, and getting that picture or sound or whatever it is means extracting the value out of that data so people can act on it, which is why they got the data in the first place! Lots of money there, because there is lots of value there.
Now take a student struggling to understand things. Having a device that's not all that expensive that can render things to them in simple visual or aural terms is worth a lot! Interacting with it? Worth even more because they then can act on that information, come to perceive things that they might struggle with otherwise, and all of that rolls up into skills and understandings they need to grow and be productive, secure work, succeed. That's the kind of "can" things Ken is looking for here. I think a P2 can do that, and there is value there.
The gaming idea is an intriguing one too. Maybe casino gambling has moved beyond the point of it making sense, but there is a lot of gaming out there, and one thing about gaming is trends. Quick to market is king! If somebody does have an idea that is in scope for P2, they are highly likely to execute quickly on it, and that's another "can", and I'm operating under the assumption that we will have the core libraries and such needed, so let's set that aside and let things cook for a while. Of course, that's just an idea and not a definitive one, but it takes a fair number of those plausible ideas to roll up into some core efforts Parallax can do to market, position and secure sales.
As for volume products, I think a big mistake here is to compare overall industry volumes and shares for public companies of size to private ones. Often, a lack of significant share relative to the big players is seen as a lack of success overall, when the truth is success really comes down to the viability and profitability of the business. A primary goal here is to make enough to keep the ecosystem running nicely, viable, and to do so in a way that gets people paid enough to add value to continue. That goal has absolutely nothing to do with share comparisons, and everything to do with whether or not the value added appeals to enough people to pay for continuing to add the value.
If we were to compare numbers, how about comparing the 6502 to an ARM. There are 6502 cores all over the place. Several of your consumer products have them, and the numbers are huge! Does that make any real sense on a strict comparison? No. Of course not. I will however, highlight the impact of figuring out what something CAN do. A 6502 CAN do a lot of things and those people selling them, or cores, or IP, have sorted out all the ways and understand where the value is and they didn't do that by working hard to understand all the CAN'T do things, which arguably exceed the number of CAN do things by a significant margin.
"ARM has a lower cost"
Well yeah. Lots of things have lower costs, but that doesn't tell the whole story. Ask your typical Apple user about lower cost and then ask them about the value they see for the dollar and I think you would find that a very interesting discussion! They will tell you that the price they paid for the value they got is a great deal, and they would tell you why and nothing about that has to do with the lowest cost. It never does. Now people don't want to pay too much, but they very often can and will pay for value for the dollar and where there is more value, there are more dollars. This plays out all over the place, and the reality is cheapest isn't always the right call. Again, it's super easy to default to that, and many do. The challenge here is figuring out who sees what value and making sure they are aware of the offerings.
So the real discussion here isn't this costs X vs that which costs Y, nor is it about this has share X as opposed to that which has share Y, because neither of those things really speaks to the value of the product we are discussing.
Really, what I see here is that you don't see a lot of value in the P2. That's OK! You don't have to, but the thread is about who does and what they might do, why they might do it and so on. At any given time it is extremely easy to present the "can't" or the "not worth" aspects of things. We are all critical in this way, and again it's super easy because nothing in the world is optimal. It is significantly harder to think about the optimal and zero in on that, but when that is done the result is a much better understanding of what can be done, why it should be done, etc... which is what Ken is asking for, and rightfully so as he's the "buy stuff" guy, and we all need a "buy stuff" guy somewhere working on the task of people buying stuff because the best ideas, engineering, products in the world don't really mean much unless we have people buying stuff.
And that's hard. It's as hard as the engineering is, and sometimes I would argue harder based on my own experiences with this, again why we are having the discussion.
There are others they offer. It's a fun music device, and it's to be used for entertainment, learning, effects, live performance, studio....
The P2 has a crazy number of analog 9 bit channels, and it's not tough to add 16 bit or 24 bit channels with a helper chip or two, leaving a P2 to move data. At audio frequencies, maybe a P2 can do some of those things too, I just don't know yet. Maybe others here do, but my point is this niche is hot and growing right now!
What would a similar product look like that featured a touch LCD of some sort where areas of the pad can be highlighted to assist in playback, or to select effects, or visualize various things? What if that device were to perform as part of a MIDI chain? What if a few of them could work together? How about that device smart enough to manage the input, doing things like staying in the key that makes sense, or quantizing input in various ways to match percussion or key off of other events, vocals?
Now, look at one of these! (and I want one of these, just because)
Audio is being explored in big ways right now. Those devices connect to a PC, lots of software, etc... Lots of people looking to explore here, build, play, do, enjoy, explore...
Anyway, that's an area that might be of interest on a few levels...
I'll keep this one a positive suggestion (kind of)
PASM is needed for people working on the bare metal, so it is somewhat a necessary evil.
But other than Bean's BASIC or the GCC project, there seems to be no effort to support a standard language. Take a look at the Parallax site and outside the forum try to find any reference to GCC or BASIC on the Prop, and even those are difficult to find.
At least on BASIC, Parallax had a huge installed base. It would have been a natural follow on to the Stamps. So why no PBASIC aimed at the P1? I mean one supported by Parallax, documented on the main page, and part of the installation package. A PBASIC could even take advantage of the multiprocessor nature, how hard would it be to explain MAIN1: MAIN2: MAIN3: ...
The same should be true for GCC, though I don't know the mechanics enough to know whether LMM is required or whether a real small package could be put together for it.
If you are pushing the P2/P1 is so easy to use, well make it easy to use. The standard package should install and support BASIC or C.
< negative - again > It is an awful hard sell to say this is a wonderful and easy to understand hardware engine where everything is orthogonal and easy to understand, and oh by the way you have to program it in assembly language or this interpreted language we invented.
The standard languages are C, C, and C with perhaps a little C++ thrown in.
Watching the propgcc forums for the last year and the conference presentations yesterday we see that a lot of effort has been and is being put into that.
Do bear in mind that all this is fairly new. PropGCC is still in beta test. Although from what Jazzed said yesterday it's a case of forgetting to go out of beta.
More importantly it seems parallax want the tools, SimpleIDE and such, and the documentation to be top notch before going wild publicly about it all.
It will all support Spin as well, cross platform for Windows, Mac and Linux.
My feeling is that BASIC is history. Now a days they are teaching kids programming starting with things like Python. Show them BASIC and they will not be impressed. This also makes me think that the whole point of BASIC as a simple teaching language was misguided in the first place. Evidence suggests you can start people off with much more sophisticated and powerful languages. They get along just fine if you do it right. Check out all those Arduino using beginners that are quite happy with C++.
At least on BASIC, Parallax had a huge installed base. It would have been a natural follow on to the Stamps. So why no PBASIC aimed at the P1? I mean one supported by Parallax, documented on the main page, and part of the installation package.
I've said it before and I'll say it again, it is hard to find much fault with this logic; simply look at how often posts asking about BASIC show up here in the forums (in toto). Like it or not, Parallax made their name with BASIC. Of course, some of that user base has likely already been lost for good.
< negative - again > It is an awful hard sell to say this is a wonderful and easy to understand hardware engine where everything is orthogonal and easy to understand, and oh by the way you have to program it in assembly language or this interpreted language we invented.
Likewise true, at least in the minds of many. You'll find the same complaint repeated over and over and over.
If you want a BASIC, or there is demand for a BASIC, one can be built on GCC, right along with most anything else. C as a foundation in this way rocks pretty hard IMHO.
As for PASM and easy. Yes, and it's brilliant, beautiful, productive, fun. Why fear assembly? IMHO, that skill needs to continue to get out there. The real wizards, who make stuff go for those using higher level languages use assembly for that purpose. And when push comes to shove, having that skill in your tool box means taking things to the absolute limit when needed. Count me as a fan.
Comments
That is exactly what I mumbled to myself 15 minutes ago.
These are a major pain for me to solder and so I'd very much like to use these little modules....
Yes of course you are so right.
96 i/o plus..... how many total?
At 0.1 inches represents about ten inches of pins.
2 times 50; 4 times 25 or about 2.8 inches square (round the edge), is HUGE!!!!
Point taken:thumb:
Dave
Yes but if you're driving a TV or monitor it makes perfect sense to have something that screws into the Vesa mounts of it, which are either on a 75x75 or 100x100 mm square pitch.
We have to wait for accurate power consumption figures, but depending on cogs required and clock rate it may even be possible to draw 2.5 watts out of the USB socket that many tvs & monitors have onboard
-Phil
Would appreciate hearing more ideas like this one. Contrary to what Chip suggests, we will have some customer-focused presentations to sell the P2. No boring PPT, but engineering-ready appnotes and just the right amount of presentation slides for the managerial types. I would enjoy hearing all of your suggestions for the best applications of the P2. Please do share.
If you've been through a casino recently you might notice that "slot machines" are quickly morphing into generic video games, with high resolution dynamic graphics. When slot machines started to be computerized in the 1980's the regulators would conduct spot checks to pop the EPROMs and put them in a comparator box to make sure they hadn't been hacked. I'm not sure what they're doing nowadays but I'm sure it can't be the same, unless there is some kind of Chinese wall between the display module probably running Linux and the gaming module still running old tech. Anyway the regulatory environment in Nevada is key to entering the industry, since Nevada requires manufacturers whose products are approved for Nevada to comply with those same requirements globally whether other venues are as strict or not.
In bars you will sometimes find games consoles on some of the tables. Again high res screens and networked for multi-players modes.
Just the other day I found one displaying some Windows "Fatal error" dialog box in the middle of the screen. I clicked the OK button and it continued to try and persuade me to play Trivial Pursuit by means of a nice looking girl in a video suggestively winking at me. Very anoying. I wished I'd left the fatal error dialog up.
The Props role in a gaming rig now a days might be in some sort of controller input device.
it reminds me a huge LED screen in Montreal where ads were displayed. One day, a big "Windows update" and restart notice popped up in the screen. shi* happens when you use Windows OS
Slot machines are most definitely networked at the level of the player's club card reader and display, but if they're made in Nevada the gaming modules are most definitely not allowed to communicate with each other in any way. This doesn't mean someone outside Nevada isn't doing that but it would be considered sleazy in traditional industry circles.
The bar machines which aren't "gambling" devices because you can't win money from them, most definitely are networked, phone home periodically, and are running LInux. I saw an article about the biggest distributor some years ago and they were quite bullish on the ability to distribute new games and monitor income without physical site visits.
I just don't know how the compliance enforcement would work on a machine with a surface mounted flash chip that can be field reprogrammed but not mounted in a comparator to see if that's been done. Unfortunately with the decline of the industry and opportunities for advantage play I lost interest in keeping track about the time the technology started to change. But if a company like Parallax wants to make an entry into the gaming world, someone would have to find out what the new rules are.
The big boys in gaming are IGT and Bally. IGT has a big operation in Reno. Last time I checked (about 5 years back) most machines had the gaming front end and a linux backend. The Linux part did all the communication, player tracking. Not sure where the P2 would fit into that, while yes it does video on the scale of retro games, can it do MP3 video decoding? Yes the P2 can do the older bar video poker games, that technology is well over 10 years old. Yes you could build a better version of that, but people designing those today are going to be using the technology available in any $100 Android tablet.
Parallax has to focus on what it has been good at, and use that. That has been the hobbyist education business. Is there some niche for the P2 out there, maybe. But the microprocessor field has always been very crowded. Look at the list of architectures that no longer exist or are so minimal in the market (Z80, 6800, 68000, 960, COPs, Alpha, 6500, 8051 and more that I've forgotten). The general market while large is coalescing around a couple architectures, ARM and maybe PICs at the real low end, though in my opinion the cost of the packaging will exceed the cost of the die and at that point the PICs will go away. We are probably less than a couple years from that with ARMs in volume well less than $1.
Chips from the LPC800 family (Cortex M0+) are already in the $0.59 range, small qty. Really just depends on which ARM you want. There certainly is a broad spectrum of performance and cost. The NXP chip I use most frequently is $1.26, single piece. I haven't even thought of using a PIC in the last 10 years! Crazy how things evolve.
Anyway, I agree that Parallax should continue to market products to the educator and hobbyist communities. But I do also think there are certain niche markets to be had in, eg, scientific and industrial applications.
Making that determination is actually a task with a complexity on par with developing the product! I've been through a couple of those exercises. It was very interesting! A lot of the problem boils down to what others see as value and coming to understand what that value is worth to them. You can get a baseline that way, and it's messy. Gotta talk to a lot of people, get data, run scenarios, produce reference designs, concepts, etc... A lot more of the problem comes down to understanding what the competing offerings are, which will indicate a variance on the price to compete. Finally, there is the risk costs. Product life window, time to market, etc... Hit early, price can be high, margins high, because early adopters are buying time and taking risks for their own reasons. Hit in the middle, price is solid, margins reasonable, etc... Late? Well, price is going to be lower, margins thinner.
Once all that is sorted out, price is somewhat known, and if that price works, do it. If not, value must be added so that price does work.
Wash, rinse, repeat over the life cycle of whatever it is.
Hey Bruce! The group of "hobby education" is two different audiences with 20-30% overlap in our experience. Educational markets are addressed in person, by training teachers.
You omitted the commercial market as a focus of ours. Did you know that we have production customers using Propeller 1 who will also design in Propeller 2? These customers are an absolute necessity to our future success and will be the continued focus of our FAEs. The latest production effort is www.swapbox.com (shared with their approval, of course). Production customers who choose use the Propeller for rapid prototype to production, flexible I/O from software, ease of multicore, object support, personal FAE assistance, and it makes their products far more practical to develop. Their applications are in renewable energy, robotics, medical, environmental measurements and machine control. We needn't sell ARM volumes to this audience to prosper, only a few million chips a year once P2 is underway.
I think I understand your point, and simply want to point out that we have relevant production customers.
But the question you have to answer, if you are starting a clean sheet design, would you really look at much other than an ARM for an embedded project? Sure the user manual is close to 1000 pages for many of them, but the difference between a hobby/education project and a project that will go into even limited production, is that a team of people working full time for a couple months gets amortized over the production run. The hobby project gets amortized over a couple units.
The ARMs will have a lower cost and are a 1 chip solution, no boot Flash or core power supply needed. While the prop can do limited specialized IO, it can't compete with dedicated hardware. Take a $8 ARM these days and compare to a P2. That ARM will do 4 UARTs, 100 Mb Ethernet, 2 CAN, 2 USBs, LCD/VGA controllers all built in and all capable of running in parallel. Now the P2 can do any couple of those, but not all simultaneously, and it still requires extra support hardware.
This is not to say there is no place for the P2, it's just unlikely to be a volume product. Maybe in the lab or education it will shine and capture enough of these to build a business. Maybe there is a market as a peripheral to be licensed to bigger vendors, though there are already homegrown versions of this at both NXP and TI.
Perhaps you do already have sales reps out there, but when I checked the website(s) I couldn't find any mention of them.
KC_Rob, no, we don't use sales representatives at the moment. I'm familiar with them from our own design-in processes when we source parts from Future, Arrow or other companies that use representatives. We are in talks with a distributor of this type who use representatives throughout the country, as this kind of approach would be more manageable than us working directly with individual representatives. We have only 42 people at Parallax so a representative arrangement would best be handled by a stocking distributor. I appreciate this thought and recognize the value of personal technical sales support.
I will say, though, that I have worked for smaller companies (less than $10 million annually) that have managed to retain and work with sales reps throughout the US on their own without too much difficulty. Once the reps are selected and contracts in place, the process becomes somewhat routine.
Those efforts will be replicated on the P2, and in fact are happening right now. The years spent serving educators positions Parallax very well to compete in education / labs with P2. I would not characterize this as a "maybe", but more like "how much?" two entirely different things! Frankly, I think some of Chip's vision for the "bench computer" where visualizing real world phenomena has some serious appeal in education and labs. Some of the very best learning happens when we can get our senses wrapped around stuff. Being able to visualize, manipulate, capture data, produce feedback, interact, control is a very high value activity! We know it's high value because just look at the cost of gear designed to do just that. Look at what people will pay for visualization, just as one related example. In that world, people take large data and they break it down to a picture. The PhD people, can look at the data and visualize internally, sometimes... most everybody else can't most of the time, and getting that picture or sound or whatever it is means extracting the value out of that data so people can act on it, which is why they got the data in the first place! Lots of money there, because there is lots of value there.
Now take a student struggling to understand things. Having a device that's not all that expensive that can render things to them in simple visual or aural terms is worth a lot! Interacting with it? Worth even more because they then can act on that information, come to perceive things that they might struggle with otherwise, and all of that rolls up into skills and understandings they need to grow and be productive, secure work, succeed. That's the kind of "can" things Ken is looking for here. I think a P2 can do that, and there is value there.
The gaming idea is an intriguing one too. Maybe casino gambling has moved beyond the point of it making sense, but there is a lot of gaming out there, and one thing about gaming is trends. Quick to market is king! If somebody does have an idea that is in scope for P2, they are highly likely to execute quickly on it, and that's another "can", and I'm operating under the assumption that we will have the core libraries and such needed, so let's set that aside and let things cook for a while. Of course, that's just an idea and not a definitive one, but it takes a fair number of those plausible ideas to roll up into some core efforts Parallax can do to market, position and secure sales.
As for volume products, I think a big mistake here is to compare overall industry volumes and shares for public companies of size to private ones. Often, a lack of significant share relative to the big players is seen as a lack of success overall, when the truth is success really comes down to the viability and profitability of the business. A primary goal here is to make enough to keep the ecosystem running nicely, viable, and to do so in a way that gets people paid enough to add value to continue. That goal has absolutely nothing to do with share comparisons, and everything to do with whether or not the value added appeals to enough people to pay for continuing to add the value.
If we were to compare numbers, how about comparing the 6502 to an ARM. There are 6502 cores all over the place. Several of your consumer products have them, and the numbers are huge! Does that make any real sense on a strict comparison? No. Of course not. I will however, highlight the impact of figuring out what something CAN do. A 6502 CAN do a lot of things and those people selling them, or cores, or IP, have sorted out all the ways and understand where the value is and they didn't do that by working hard to understand all the CAN'T do things, which arguably exceed the number of CAN do things by a significant margin.
"ARM has a lower cost"
Well yeah. Lots of things have lower costs, but that doesn't tell the whole story. Ask your typical Apple user about lower cost and then ask them about the value they see for the dollar and I think you would find that a very interesting discussion! They will tell you that the price they paid for the value they got is a great deal, and they would tell you why and nothing about that has to do with the lowest cost. It never does. Now people don't want to pay too much, but they very often can and will pay for value for the dollar and where there is more value, there are more dollars. This plays out all over the place, and the reality is cheapest isn't always the right call. Again, it's super easy to default to that, and many do. The challenge here is figuring out who sees what value and making sure they are aware of the offerings.
So the real discussion here isn't this costs X vs that which costs Y, nor is it about this has share X as opposed to that which has share Y, because neither of those things really speaks to the value of the product we are discussing.
Really, what I see here is that you don't see a lot of value in the P2. That's OK! You don't have to, but the thread is about who does and what they might do, why they might do it and so on. At any given time it is extremely easy to present the "can't" or the "not worth" aspects of things. We are all critical in this way, and again it's super easy because nothing in the world is optimal. It is significantly harder to think about the optimal and zero in on that, but when that is done the result is a much better understanding of what can be done, why it should be done, etc... which is what Ken is asking for, and rightfully so as he's the "buy stuff" guy, and we all need a "buy stuff" guy somewhere working on the task of people buying stuff because the best ideas, engineering, products in the world don't really mean much unless we have people buying stuff.
And that's hard. It's as hard as the engineering is, and sometimes I would argue harder based on my own experiences with this, again why we are having the discussion.
http://shop.korgusa.com/p/kaossilator?pp=6
And a sample of how it's used: https://www.youtube.com/watch?v=aeQOuNBuJwg
There are others they offer. It's a fun music device, and it's to be used for entertainment, learning, effects, live performance, studio....
The P2 has a crazy number of analog 9 bit channels, and it's not tough to add 16 bit or 24 bit channels with a helper chip or two, leaving a P2 to move data. At audio frequencies, maybe a P2 can do some of those things too, I just don't know yet. Maybe others here do, but my point is this niche is hot and growing right now!
What would a similar product look like that featured a touch LCD of some sort where areas of the pad can be highlighted to assist in playback, or to select effects, or visualize various things? What if that device were to perform as part of a MIDI chain? What if a few of them could work together? How about that device smart enough to manage the input, doing things like staying in the key that makes sense, or quantizing input in various ways to match percussion or key off of other events, vocals?
Now, look at one of these! (and I want one of these, just because)
http://www.youtube.com/watch?v=3vC5TsSyNjU
Or a more sane example:
http://www.youtube.com/watch?v=y1BngLOA9Z4
Audio is being explored in big ways right now. Those devices connect to a PC, lots of software, etc... Lots of people looking to explore here, build, play, do, enjoy, explore...
Anyway, that's an area that might be of interest on a few levels...
I have always wondered about Prop II and audio applications. Basically because the Prop II's nearest rival has already been there for a while. Dare I say it, XMOS.
For example they are seriously into USB audio gizmos, for example this:
http://www.ebay.com/itm/XMOS-384kHz-high-quality-USB-to-I2S-PCB-with-ultralow-noise-6-5uV-regulator-/230961204617?pt=US_Amplifier_Parts_Components&hash=item35c65c4d89
and also see
http://www.xmos.com/discover/possibilities/audio
PASM is needed for people working on the bare metal, so it is somewhat a necessary evil.
But other than Bean's BASIC or the GCC project, there seems to be no effort to support a standard language. Take a look at the Parallax site and outside the forum try to find any reference to GCC or BASIC on the Prop, and even those are difficult to find.
At least on BASIC, Parallax had a huge installed base. It would have been a natural follow on to the Stamps. So why no PBASIC aimed at the P1? I mean one supported by Parallax, documented on the main page, and part of the installation package. A PBASIC could even take advantage of the multiprocessor nature, how hard would it be to explain MAIN1: MAIN2: MAIN3: ...
The same should be true for GCC, though I don't know the mechanics enough to know whether LMM is required or whether a real small package could be put together for it.
If you are pushing the P2/P1 is so easy to use, well make it easy to use. The standard package should install and support BASIC or C.
< negative - again > It is an awful hard sell to say this is a wonderful and easy to understand hardware engine where everything is orthogonal and easy to understand, and oh by the way you have to program it in assembly language or this interpreted language we invented.
https://www.youtube.com/watch?feature=player_embedded&v=822F2VQAKF4#!
The standard languages are C, C, and C with perhaps a little C++ thrown in.
Watching the propgcc forums for the last year and the conference presentations yesterday we see that a lot of effort has been and is being put into that.
Do bear in mind that all this is fairly new. PropGCC is still in beta test. Although from what Jazzed said yesterday it's a case of forgetting to go out of beta.
More importantly it seems parallax want the tools, SimpleIDE and such, and the documentation to be top notch before going wild publicly about it all.
It will all support Spin as well, cross platform for Windows, Mac and Linux.
My feeling is that BASIC is history. Now a days they are teaching kids programming starting with things like Python. Show them BASIC and they will not be impressed. This also makes me think that the whole point of BASIC as a simple teaching language was misguided in the first place. Evidence suggests you can start people off with much more sophisticated and powerful languages. They get along just fine if you do it right. Check out all those Arduino using beginners that are quite happy with C++.
Likewise true, at least in the minds of many. You'll find the same complaint repeated over and over and over.
As for PASM and easy. Yes, and it's brilliant, beautiful, productive, fun. Why fear assembly? IMHO, that skill needs to continue to get out there. The real wizards, who make stuff go for those using higher level languages use assembly for that purpose. And when push comes to shove, having that skill in your tool box means taking things to the absolute limit when needed. Count me as a fan.