To whatever extent my comments are perceived as rude, I apologize. I felt it necessary to express myself frankly, because I wasn't sure the message was getting through.
Featuritis can be an addiction. And when we get together with our featuraholic buddies to talk about more features, it creates an endless, self-sustaining cycle, not unlike those of other addictions. The danger is that the activity becomes an end in itself. What I'm trying to say is that it's time to sober up, not necessarily that the features that have been added are bad. It's just that maybe feature Q is a good idea and will make the P2 better than what we have. But once that's added how about R, S, T, U, ... ? Well you get the idea. None of those features count for anything if there's no physical IC to manifest them. And there won't be as long as the door remains ajar for yet more suggestions. And, as a corollary, the P2, when completed, will never be as good as it might have been if the dev cycle had been extended just a little while longer. See what I mean?
I'm just as prone as anyone to confuse activity with accomplishment. I do it all the time. But then I wake up hung over and think to myself, "Migosh, this has been going on now for how many years?" Maybe it's time to sober up and get with the twelve-step program to recovery.
"Hi, my name is Phil, and I'm a featuraholic."
Phil, I appreciate what you are saying. I've been working on this project for 8 years now and in the last several months, things have sped up quite a bit, so I'm packing what I can into it before we send it off for fab again. The cost of these few months will, I'm confident, pay off later, as we've got a chip that will be more amenable to the C juggernaut that you and I don't care for so much, but that the world expects - at least as an entry point. We can get good use out of the C-friendly features, too.
The cost of these few months will, I'm confident, pay off later, as we've got a chip that will be more amenable to the C juggernaut that you and I don't care for so much, ...
'Spent another hour today helping H.S. students get their ActivityBots running with C. I just kept thinking how much easier it would be to teach them Spin and how much more we could accomplish in the limited time available. But they wanted C, and sometimes we just have to hold our noses and satisfy the demand that exists.
Phil, I appreciate what you are saying. I've been working on this project for 8 years now and in the last several months, things have sped up quite a bit, so I'm packing what I can into it before we send it off for fab again. The cost of these few months will, I'm confident, pay off later, as we've got a chip that will be more amenable to the C juggernaut that you and I don't care for so much, but that the world expects - at least as an entry point. We can get good use out of the C-friendly features, too.
I suspect that the features you've added to support C would be equally useful for any language that compiles to native code including Spin if anyone ever writes a native code compiler for it.
To whatever extent my comments are perceived as rude, I apologize. I felt it necessary to express myself frankly, because I wasn't sure the message was getting through.
I think you've been fairly frank several times. Perhaps a failure to get your opinion over isn't the problem. Could just be that others are of a different mind. Perhaps not all the P2 suggestions that have been offered are motivated by pure selfishness and greed, either.
Tubular, heater, and others have made outstanding points that are quite persuasive to me.
I think you've been fairly frank several times. Perhaps a failure to get your opinion over isn't the problem. Could just be that others are of a different mind. Perhaps not all the suggestions that have been offered are motivated by pure selfishness and greed, either.
Tubular, heater, and others have made outstanding points that are quite persuasive to me.
True. The requests that various of us have made for features to support C/C++ are not out of selfishness. The PropGCC project was started by Parallax and funded by them. It is a priority to them and making it run efficiently on the P2 serves to improve their chances of success with P2. Also, I'd be willing to bet that Catalina C will benefit from these changes as well.
I have been timid and I have stated my observations while trying to placate the party goers, Phil has been a bit braver and stepped in and said what I am thinking. The result? It's just like Jimmy Kimmel's prank and some have become very upset, some almost abusive. Of course I want all those features you dream up and perhaps some of my own ideas too, especially those that support Forth implementation But that's a dangerous road to go down because it's hard to stop (the Forth one too )
All those features are not even hot air because that would have some substance at least, emulation or not. How many of us would be here if we didn't have P1 silicon? Would we be conversing with Chip about all these great ideas for P1 after all these years? Would we know who Chip is? Well in effect the P1 if it wasn't silicon yet and somehow managed to have those featuraholics contributing would actually be a P2 before it was a P1 (as we know it in silicon). This is what has happened to the P2
Now I see the featuraholics getting very emotional as they are confronted with their own problem and choose to lash out instead. Phil deserves a lot of respect as a long time supporter, contributor, and moderator of this forum. Follow Chip's lead as he does just that and also speaks rationally about the project which certainly instills confidence in me that he has a handle on it at least. I've been worried that the long hours, delays, workload, business pressures etc have made him a junkie too!
All of us are in this together, we all have Propeller caps on, we are all a bit weird, we all want this to work. No need to shout down a voice, one of us, just listen and think about it.
As I have said before I want a piece of the P2 pie but I will take it as it is, eat it for breakfast lunch and dinner, then maybe look to P3 for more. Intel didn't jump from the 8086 to the Pentium but this is the kind of design gap between the P1 and P2. Intel is still around and doing very well.
Question. Did Phil really say that these "features" were being requested out of greed and selfishness? Maybe it's somewhere else, but I assume this is what is being referenced when the accusation is made:
These guys are not your friends, nor the friends of a company and its employees that hope to profit from your efforts. They're hobbyists and, at best, low-volume users. What do your potential OEM customers want that other semiconductor companies are not providing? Do you have any idea?
Now maybe that is what he meant - ie, people requesting features are doing so for greed and profit - but that is not in fact what he wrote above. Or did he write something else which says that more explicitly?
I think all the concerns about delays are due to the fact that the Propeller 2 has become much more visible recently, and not due to any new problem.
Why is this true? Because almost a third of all Propeller 2 discussion has happened in the last month!. That's pretty cool. Chip finally has a bunch of people talking over the ideas in a sustained conversation.
What about threads? It seems that every day there is a new Propeller 2 thread.
All these posts and threads makes it seem like the Propeller 2 is taking forever to come out, but in my mind Parallax and Chip haven't been working on the Propeller 2 for the last 8 years. They've been preparing the groundwork in order to have this month long discussion on what would make the best design.
Data for these images was collected by scanning the Propeller 2 forums for posts before and after 2013-11-11 (one month ago) and adding up the replies to posts. If a thread was started before 11-11 it was put into the before bucket, regardless if it had any posts later. The only exception is the Propeller II Update - Blog thread, which was split into two counts for before and after.
Question. Did Phil really say that these "features" were being requested out of greed and selfishness? Maybe it's somewhere else, but I assume this is what is being referenced when the accusation is made:
Now maybe that is what he meant - ie, people requesting features are doing so for greed and profit - but that is not in fact what he wrote above. Or did he write something else which says that more explicitly?
I think it was a reference to the paragraph before the one you quote.
I hate to say it, but what you've gotten are a bunch of sycophants who say they love you for entertaining their wild fantasies, but will they respect you in the morning?
Question. Did Phil really say that these "features" were being requested out of greed and selfishness?
There are no quotes around those terms on purpose. Here are some quotes:
...but what you've gotten are a bunch of sycophants who say they love you for entertaining their wild fantasies, but will they respect you in the morning? Unless you deliver a finished product -- and soon -- I doubt it. Moreover, if you expect Parallax to profit from the P2, these are not the people you need to be listening to.
I guess it's a question of whose self-interests are being served by this process.
We are all free to draw our own conclusions. I've drawn mine.
True enough. Although it accomplished my objective of getting people's attention and spurring discussion, the same end might have been attained with less inflammatory language. Still, though, from a metphoical standpoint, I thought that sentence forged some rather vivid imagery! Besides, who amongst us is ready to claim that any suggestions we've made for the P2 are entirely selfless and altruistic? Certainly not I! It's like, "C'mon, Chip, have one more round with us, then you can go home. We promise."
(I'm beginning to feel like an author whose book is being taught at college level and imbued by the prof at every turn of phrase with significance that never occurred to me. But keep up the analysis; I'm rather enjoying it now! )
Finally, it is of course possible that I'm repeating the fallacy of Zeno's paradox, where Chip is Achilles, and a finished P2 is the tortoise.
True enough. Although it accomplished my objective of getting people's attention and spurring discussion, the same end might have been attained with less inflammatory language. Still, though, from a metphoical standpoint, I thought that sentence forged some rather vivid imagery! Besides, who amongst us is ready to claim that any suggestions we've made for the P2 are entirely altruistic? Certainly not I! It's like, "C'mon, Chip, have one more round with us, then you can go home. We promise."
(I'm beginning to feel like an author whose book is being taught at college level and imbued by the prof at every turn of phrase with significance that never occurred to me. But keep up the analysis; I'm rather enjoying it now! )
Finally, it is of course possible that I'm repeating the fallacy of Zeno's paradox, where Chip is Achilles, and a finished P2 is the tortoise.
-Phil
Actually, I agree with your concern about features being piled onto P2. I think I made a comment to that effect myself a while back. But when Chip mentioned that he was going to look at executing code from hub memory I couldn't help myself! I never thought he'd really consider that. Even at this point, I'd understand if it turns out not to be feasible but I do think it will improve C/C++ performance and is worth doing if it isn't a huge risk.
We are all free to draw our own conclusions. I've drawn mine.
Of course you can. But greed, as commonly used and understood, isn't there, implicitly or explicitly. Read complete and in context, probably Phil's ad hominem charge is that of a fanboi-driven toadyism. But take your pick.
Still, though, from a metphoical standpoint, I thought that sentence forged some rather vivid imagery! Besides, who amongst us is ready to claim that any suggestions we've made for the P2 are entirely selfless and altruistic?
I understand now and had, I think, a pretty good idea of what you were getting at. We all, whether we admit it or not, operate in terms of self-interest. That's just a fact of life. (However, that does not always mean for money or material wealth - ie, greed.)
Nonetheless, I myself prefer to stay away from the personal as much as possible in these discussions. I've called out others, with views the opposite of yours, for that very thing. So it would be unprincipled of me not to say that your language was a little over the top there.
Even at this point, I'd understand if it turns out not to be feasible but I do think it will improve C/C++ performance and is worth doing if it isn't a huge risk.
At least you're honest and acknowledge that there is risk. There are tradeoffs. Eg, hours spent changing are hours that could have been spent checking, verifying, or documenting (or a dozen other tasks). Yet one can accept this and reasonably expect that the cost is worthwhile. Or not. It's a reasonable thought process either way.
I think all the concerns about delays are due to the fact that the Propeller 2 has become much more visible recently, and not due to any new problem.
Why is this true? Because almost a third of all Propeller 2 discussion has happened in the last month!. That's pretty cool. Chip finally has a bunch of people talking over the ideas in a sustained conversation.
What about threads? It seems that every day there is a new Propeller 2 thread.
All these posts and threads makes it seem like the Propeller 2 is taking forever to come out, but in my mind Parallax and Chip haven't been working on the Propeller 2 for the last 8 years. They've been preparing the groundwork in order to have this month long discussion on what would make the best design.
Data for these images was collected by scanning the Propeller 2 forums for posts before and after 2013-11-11 (one month ago) and adding up the replies to posts. If a thread was started before 11-11 it was put into the before bucket, regardless if it had any posts later. The only exception is the Propeller II Update - Blog thread, which was split into two counts for before and after.
Wow! This helps put things in perspective. Thanks for getting those graphs together. I notice nowadays that there might be 45 people on the Prop2 forum at once, which is a huge increase over past numbers.
Those graphs just reinforce what I already knew.
The start to this huge improvement round was with Chip's release of the new instruction set. This was shortly followed by the Thanksgiving holidays when Chip asked about DDR2. While this was not a goer, it spurred a lot of understanding about P2s timing and bandwidth and this resulted in some very serious advances in the P2, such as the space achieved by the DAC bus reduction that gave us 256KB Hub and WIDE (8 Long) accesses, and then the discussions about Hubexec mode.
We are all going to be significantly better off for these dramatic improvements in the P2, and that in turn hopefully opens up a lot more commercial avenues for the P2 to excel in.
Yeah, I'm with Cluso. Took me a while to see it, but we are on a great path right now. I've some worries, but nothing like they were before.
@All: I saw Phil's point, and share a couple of minor worries. I too did not mean to imply greed or something nefarious regarding contributors to P2. If anything, I was speaking to bias that might not lead to optimal design choices, but not any bad intent on the part of anybody. Those worries are from more standard experiences, and sometimes they are hard to shake. This project isn't standard at all, and sometimes people may find it difficult to see it for what it is. I have.
@Chip: And to be really clear, I see you are feeding from the dialog. Honestly, working in a bubble, somewhat alone, not connected is hard. Lots of good energy going on right now. This is good to see, and that is what I meant above.
Right now, I believe this is going to be a great chip. From time to time, I've mentioned "the Apple of micro-controllers" and I mean that in a very good way. Apple expanded and to a degree redefined the portable media player market. Early adopters in that market actually turned right around and bought in twice! I'm not really sure those dynamics are easily compared to this market, but I think the general idea is worth aiming for. The P2 is going to flat out SING in PASM + SPIN.
And this time we've got solid support for C, which is going to pack a very nice punch once a good set of libraries are built up. My early experiences with gcc, when we had it up on the first FPGA, were very favorable. Didn't feel at all like the P1. And that's the right path. Has to be.
On the educational front, there is enough room and there will be modern, relevant options that can present the tech in great ways for students. Personally, I think this will very seriously enhance Parallax education, and that's a known quantity filled with many high probability adopters as well as prospects. IMHO, the single biggest initial market there is, and the features are well aligned with it too. I'm not really sure this could be better than it's shaping up to be right now.
For those wanting to learn in a less structured way, SPIN 2 including PASM snippets is going to be fantastic. One of the major barriers to learning any assembly language is context. Everybody seems to be able to grasp small bits of code easily. That is in terms of what it is doing. However, doing that with enough context to do something useful takes a lot more. I think people are going to be able to write in SPIN, then begin to author little, lean bits of PASM to do specific things, with a nice framework in place. When many of us learned assembly language, if we were not doing machine language due to lack of tools, we were very often starting out with little routines to boost up BASIC, or access a hardware feature not exposed, etc... Those little bits of code mean learning how assembly language works, and it means getting the benefit of it without also having to build up a new tool chain, or the larger structure required for more meaningful programs at the same time. Tutorials can be written to perform a task in spin, contain easy to understand and fun inputs and outputs beyond the blinking light or toggling pin, and the tests needed to verify proper function. People can complete one "fill in the blanks" style and use the output from the tests, along with aural, graphical, I/O pin feedback, a trace run, etc... to understand what is happening.
Finally, there is enough speed and storage options to allow for fairly painless big programs. Everybody will benefit from this, and the overall scope of possible and practical things goes up a huge amount too. Being able to build up "full" applications without needing an operating system is going to be notable and it's lean by definition. I look forward to seeing how this plays out too.
Speed and I/O is one thing, just having room to afford some luxery is going to be very appealing, not only compared to P1, but to other devices.
Well, really finally, it's also going to be the "Dodge" of micro-controllers. Ever notice how Dodge regularly says, "Best in class?" Well, they get to do that largely because their vehicles do not fit precisely into the more established classes, which makes direct comparisons more difficult, and they've done a fine job of turning that into a positive, with "best in class" type messages that can resonate. P2 will have this general property about it. There won't really be a similar class, and I think that is overall a very good thing. Highly differentiated products have some advantages in this way. Let's hope they play out favorably.
There is competition from companies like Freescale and TI not just the "the company that cannot be named". They aren't the P2 but they are legit multicore offerings from major corporations. If you look these companies are investing serious resources into their multicore division.
Multicore is the new black so to speak and will only get bigger from here on out.
Wow! This helps put things in perspective. Thanks for getting those graphs together. I notice nowadays that there might be 45 people on the Prop2 forum at once, which is a huge increase over past numbers.
That's what they said about the Edsel, isn't it?
All that market research that Phil wants but the most famous flop in auto history.
This post comes to you from Jamie Oliver's in St. Petersburg.
The market research indicates that more, cheaper ARM chips are needed, along with cheaper programmers in far-away places who will handle the niggling details so that the project manager can enjoy a beer.
The market research indicates that more, cheaper ARM chips are needed, along with cheaper programmers in far-away places who will handle the niggling details so that the project manager can enjoy a beer.
I'm painfully aware of what you say about cheap programmers in far-away places. I think I'm the last software engineer left in the US who works for the STMicroelectronics R&D organization. All the real technical work is done elsewhere including India, Singapore, Tunisia, and even France and Italy. I'm stuck with the role of "software achitect" which means I write lists of product features and write PowerPoint presentations and never code. Anyone want to hire a software engineer from New Hampshire? :-)
Comments
Phil, I appreciate what you are saying. I've been working on this project for 8 years now and in the last several months, things have sped up quite a bit, so I'm packing what I can into it before we send it off for fab again. The cost of these few months will, I'm confident, pay off later, as we've got a chip that will be more amenable to the C juggernaut that you and I don't care for so much, but that the world expects - at least as an entry point. We can get good use out of the C-friendly features, too.
'Spent another hour today helping H.S. students get their ActivityBots running with C. I just kept thinking how much easier it would be to teach them Spin and how much more we could accomplish in the limited time available. But they wanted C, and sometimes we just have to hold our noses and satisfy the demand that exists.
-Phil
I think you've been fairly frank several times. Perhaps a failure to get your opinion over isn't the problem. Could just be that others are of a different mind. Perhaps not all the P2 suggestions that have been offered are motivated by pure selfishness and greed, either.
Tubular, heater, and others have made outstanding points that are quite persuasive to me.
All those features are not even hot air because that would have some substance at least, emulation or not. How many of us would be here if we didn't have P1 silicon? Would we be conversing with Chip about all these great ideas for P1 after all these years? Would we know who Chip is? Well in effect the P1 if it wasn't silicon yet and somehow managed to have those featuraholics contributing would actually be a P2 before it was a P1 (as we know it in silicon). This is what has happened to the P2
Now I see the featuraholics getting very emotional as they are confronted with their own problem and choose to lash out instead. Phil deserves a lot of respect as a long time supporter, contributor, and moderator of this forum. Follow Chip's lead as he does just that and also speaks rationally about the project which certainly instills confidence in me that he has a handle on it at least. I've been worried that the long hours, delays, workload, business pressures etc have made him a junkie too!
All of us are in this together, we all have Propeller caps on, we are all a bit weird, we all want this to work. No need to shout down a voice, one of us, just listen and think about it.
As I have said before I want a piece of the P2 pie but I will take it as it is, eat it for breakfast lunch and dinner, then maybe look to P3 for more. Intel didn't jump from the 8086 to the Pentium but this is the kind of design gap between the P1 and P2. Intel is still around and doing very well.
Why is this true? Because almost a third of all Propeller 2 discussion has happened in the last month!. That's pretty cool. Chip finally has a bunch of people talking over the ideas in a sustained conversation.
What about threads? It seems that every day there is a new Propeller 2 thread.
All these posts and threads makes it seem like the Propeller 2 is taking forever to come out, but in my mind Parallax and Chip haven't been working on the Propeller 2 for the last 8 years. They've been preparing the groundwork in order to have this month long discussion on what would make the best design.
Data for these images was collected by scanning the Propeller 2 forums for posts before and after 2013-11-11 (one month ago) and adding up the replies to posts. If a thread was started before 11-11 it was put into the before bucket, regardless if it had any posts later. The only exception is the Propeller II Update - Blog thread, which was split into two counts for before and after.
There are no quotes around those terms on purpose. Here are some quotes:
We are all free to draw our own conclusions. I've drawn mine.
(I'm beginning to feel like an author whose book is being taught at college level and imbued by the prof at every turn of phrase with significance that never occurred to me. But keep up the analysis; I'm rather enjoying it now!
Finally, it is of course possible that I'm repeating the fallacy of Zeno's paradox, where Chip is Achilles, and a finished P2 is the tortoise.
-Phil
Nonetheless, I myself prefer to stay away from the personal as much as possible in these discussions. I've called out others, with views the opposite of yours, for that very thing. So it would be unprincipled of me not to say that your language was a little over the top there.
Would that be Hypno-toadyism ?
I'm very glad Parallax wants to compete.
Wow! This helps put things in perspective. Thanks for getting those graphs together. I notice nowadays that there might be 45 people on the Prop2 forum at once, which is a huge increase over past numbers.
The start to this huge improvement round was with Chip's release of the new instruction set. This was shortly followed by the Thanksgiving holidays when Chip asked about DDR2. While this was not a goer, it spurred a lot of understanding about P2s timing and bandwidth and this resulted in some very serious advances in the P2, such as the space achieved by the DAC bus reduction that gave us 256KB Hub and WIDE (8 Long) accesses, and then the discussions about Hubexec mode.
We are all going to be significantly better off for these dramatic improvements in the P2, and that in turn hopefully opens up a lot more commercial avenues for the P2 to excel in.
Yeah, I'm with Cluso. Took me a while to see it, but we are on a great path right now. I've some worries, but nothing like they were before.
@All: I saw Phil's point, and share a couple of minor worries. I too did not mean to imply greed or something nefarious regarding contributors to P2. If anything, I was speaking to bias that might not lead to optimal design choices, but not any bad intent on the part of anybody. Those worries are from more standard experiences, and sometimes they are hard to shake. This project isn't standard at all, and sometimes people may find it difficult to see it for what it is. I have.
@Chip: And to be really clear, I see you are feeding from the dialog. Honestly, working in a bubble, somewhat alone, not connected is hard. Lots of good energy going on right now. This is good to see, and that is what I meant above.
Right now, I believe this is going to be a great chip. From time to time, I've mentioned "the Apple of micro-controllers" and I mean that in a very good way. Apple expanded and to a degree redefined the portable media player market. Early adopters in that market actually turned right around and bought in twice! I'm not really sure those dynamics are easily compared to this market, but I think the general idea is worth aiming for. The P2 is going to flat out SING in PASM + SPIN.
And this time we've got solid support for C, which is going to pack a very nice punch once a good set of libraries are built up. My early experiences with gcc, when we had it up on the first FPGA, were very favorable. Didn't feel at all like the P1. And that's the right path. Has to be.
On the educational front, there is enough room and there will be modern, relevant options that can present the tech in great ways for students. Personally, I think this will very seriously enhance Parallax education, and that's a known quantity filled with many high probability adopters as well as prospects. IMHO, the single biggest initial market there is, and the features are well aligned with it too. I'm not really sure this could be better than it's shaping up to be right now.
For those wanting to learn in a less structured way, SPIN 2 including PASM snippets is going to be fantastic. One of the major barriers to learning any assembly language is context. Everybody seems to be able to grasp small bits of code easily. That is in terms of what it is doing. However, doing that with enough context to do something useful takes a lot more. I think people are going to be able to write in SPIN, then begin to author little, lean bits of PASM to do specific things, with a nice framework in place. When many of us learned assembly language, if we were not doing machine language due to lack of tools, we were very often starting out with little routines to boost up BASIC, or access a hardware feature not exposed, etc... Those little bits of code mean learning how assembly language works, and it means getting the benefit of it without also having to build up a new tool chain, or the larger structure required for more meaningful programs at the same time. Tutorials can be written to perform a task in spin, contain easy to understand and fun inputs and outputs beyond the blinking light or toggling pin, and the tests needed to verify proper function. People can complete one "fill in the blanks" style and use the output from the tests, along with aural, graphical, I/O pin feedback, a trace run, etc... to understand what is happening.
Finally, there is enough speed and storage options to allow for fairly painless big programs. Everybody will benefit from this, and the overall scope of possible and practical things goes up a huge amount too. Being able to build up "full" applications without needing an operating system is going to be notable and it's lean by definition. I look forward to seeing how this plays out too.
Speed and I/O is one thing, just having room to afford some luxery is going to be very appealing, not only compared to P1, but to other devices.
Well, really finally, it's also going to be the "Dodge" of micro-controllers. Ever notice how Dodge regularly says, "Best in class?" Well, they get to do that largely because their vehicles do not fit precisely into the more established classes, which makes direct comparisons more difficult, and they've done a fine job of turning that into a positive, with "best in class" type messages that can resonate. P2 will have this general property about it. There won't really be a similar class, and I think that is overall a very good thing. Highly differentiated products have some advantages in this way. Let's hope they play out favorably.
Yes the landscape changes daily.
That parallel processor you linked to is on a totally different planet though. Not at all compareable to a Prop.
Well, no
I mean glad Parallax "wants to compete" in general terms. One could say that there is no competition, but that's silly (and would be toady).
Hypno-toadyism borrows from Futurama of course. http://futurama.wikia.com/wiki/Hypnotoad
Multicore is the new black so to speak and will only get bigger from here on out.
"If you build it, they will come." : ]
All that market research that Phil wants but the most famous flop in auto history.
This post comes to you from Jamie Oliver's in St. Petersburg.
The market research indicates that more, cheaper ARM chips are needed, along with cheaper programmers in far-away places who will handle the niggling details so that the project manager can enjoy a beer.