Interesting top 10 thread views - where are we now?
Peter Jakacki
Posts: 10,193
I noticed my Tachyon thread has hit the magic 200,000 views (and 1,824 replies). However if you click on views in Replies/Views on the forum title bar it's interesting which threads top the list.
First off there is my Tachyon thread which has been running since July 2012 but here's a quick summary of the top 10 (not counting any P2 threads)
1. TACHYON O/S - Fast, Forthwrite n Furiously Interactive n Compact, FAT32+Networking (Peter Jakacki)
2. Propeller GUI touchscreen and full color display (Dr_Acula)
3. Download PropBASIC here... 00.01.14 LAST VERSION FOR BST (Bean)
4. Mac/Linux/Windows IDE - Ver 0.19.3 "Now with new improved PropBasic" release (BradC)
5. ZiCog a Zilog Z80 emulator in 1 Cog (heater)
6. DEMO: IR Heartbeat detector (Beau Schwabe)
7. SPIN CODE EXAMPLES FOR THE BEGINNER (Public Version) (Dave Scanlan) very old thread though
8. Quad/Hexa-Copter using Propeller, Gyros, Accelerometers, Compass, Pressure & GP (Cluso99)
9. Propeller Assembly for beginners (Harprit)
10. Zog - A ZPU processor core for the Prop + GNU C, C++ and FORTRAN.Now replaces S (heater)
It seems the glory days of the Propeller have passed us by, there were once a plethora of Prop projects which have now dwindled to a dearth thereof. I remain optimistic that our good work can migrate to the P2 when that day comes. Ah, if only we had a version of the P2 in silicon rather than the "P3" dream that was being pushed.
First off there is my Tachyon thread which has been running since July 2012 but here's a quick summary of the top 10 (not counting any P2 threads)
1. TACHYON O/S - Fast, Forthwrite n Furiously Interactive n Compact, FAT32+Networking (Peter Jakacki)
2. Propeller GUI touchscreen and full color display (Dr_Acula)
3. Download PropBASIC here... 00.01.14 LAST VERSION FOR BST (Bean)
4. Mac/Linux/Windows IDE - Ver 0.19.3 "Now with new improved PropBasic" release (BradC)
5. ZiCog a Zilog Z80 emulator in 1 Cog (heater)
6. DEMO: IR Heartbeat detector (Beau Schwabe)
7. SPIN CODE EXAMPLES FOR THE BEGINNER (Public Version) (Dave Scanlan) very old thread though
8. Quad/Hexa-Copter using Propeller, Gyros, Accelerometers, Compass, Pressure & GP (Cluso99)
9. Propeller Assembly for beginners (Harprit)
10. Zog - A ZPU processor core for the Prop + GNU C, C++ and FORTRAN.Now replaces S (heater)
It seems the glory days of the Propeller have passed us by, there were once a plethora of Prop projects which have now dwindled to a dearth thereof. I remain optimistic that our good work can migrate to the P2 when that day comes. Ah, if only we had a version of the P2 in silicon rather than the "P3" dream that was being pushed.
Comments
But, yes that is a bit of a sorry list.
BradC has abandoned us and BST is now on death row.
My ZiCog is lying in a coma. One of the big motivators for the Z80 emulator on the Prop was the thought that "This will be really nice on the P2 where we won't need the complications of external RAM to run a Z80 system emulator. We can have the whole CP/M computer including video for the VT100 terminal in one chip!" Sadly the P2 did not materialize, interest waned, other things came up to divert my attention. When the P2 comes out ZiCog may get resuscitated.
I am surprised my Zog ZPU virtual machine makes it to the list at all. Still, it does have historic significance, it was the first time code compiled with GCC was run on a Propeller. It was the second system that could compile C that could be run on the Propeller (BDSC for ZiCog was the first and still the only C compiler that actually runs on the Prop). Apparently some simple ideas from the Zog loader were carried forward into prop-gcc.
Still, how representative of Propeller interest and activity is that top ten list? Most of the quantity of posts and views can be attributed to the hyperactivity of the project builder at the time. What if there are bazillions of smaller projects and topics that people are pursuing that over all dwarf the work of the top ten?
I think there is still quite a future for Propeller 1, and that people just need to be educated of it's existence and benefits, but I could be wrong. However, when the P2 arrives, then perhaps the P1 will have lost all of it's appeal.
Ray
It's like the 555 timer. Not the best at being a timer, not the fastest, or lowest power consumption or whatever. But very simple to use, very flexible, very usable in many varied situations, very dependable.
All it needs is for the P1 to me manufactured by many different companies and we would have the 555 timer of the MCU world.
Thinking about it now. Why is there no 555 timer of the MCU world? Sure we have billions of cheap PICs and AVRs etc around. But they are single sourced, all different with their own instruction sets and peripherals. It's mess. Why no basic MCU made by many manufacturers? We almost had that with various versions of Z80, 68xx chips etc but none of them are identical either. We almost have it with the ARM but again every ARM is different.
It has alway been my belief, that the Propeller chip is well suited for small CNC equipment, such as 3D printers, routers, drilling machines, etc..., with the exception that a few more pins would be very helpful.
What we need is some very good examples and videos of the Propeller being used for this type of application, then I believe there will be a resurgence in forum activity, from a variety of builders. However, I also believe that in order to be truly successful in that type of application, some basics will need to be established, such as motor operation, user interface, file access, etc...
I have a vague feeling I can improve that design further too - it has lots of discrete 74xx chips and there must be a solution nowadays with a second propeller, or a fpga, or a fpga running a propeller emulation. Can we dream about a prop chip with 100 I/O pins and a megabyte of internal memory...
That brings up another point....
Although there has been several threads on the use of multiple props, I personally would like to see more threads pertaining to advanced solutions provided by multiple props, with the addition of advanced interaction between two or more props, sharing resources, communication, wiring examples, etc...
For instance, there are several articles about serial communication between multiple props, but what I would like to see is multiple props communicating without serial communication, for example the monitoring of tied pins going high or low, then do this...
I hope that the P1 still goes strong if/when the P2 ever gets released. It's more than fast and capable enough for a huge number of things. Just look at all that's being done with it right now. There's no reason that it can't keep on doing exactly those things. Among the things going for it is the price and it's very low current draw as well as coming in a 40pin DIP. At my age my eyes aren't good enough to solder those tiny little thousand leggers anymore!
It's hard enough to try to convince people to learn a language, it's near impossible when they have to learn a different IDE on top of it.
Bean
I hope so to.
681398 - http://forums.parallax.com/showthread.php/125543-Propeller-II-update-BLOG
170851 - http://forums.parallax.com/showthread.php/155132-The-New-16-Cog-512KB-64-analog-I-O-Propeller-Chip
If we add in all of the other P2 threads the combined total views is probably over a million! I expect (hope) that the P2 activity will pick up by June based on hints that Chip and Ken have given.
Meanwhile, I use my P1 skills on little tasks and it works great.
I feel good times coming this year.
....and yet it is #3 on the list ? Basic is far from dying off..
Surely it is not rocket science to add IDE support, - adding a Batch Build & user highlighters allows almost any language to be supported, and Programmer Editors have no problems with this.
If an IDE supports Spin and C and PASM, then adding a 4th has to be close to trivial ?
I also like FreeBasic, for being small, predictable and nimble. Very good for command line Comms examples and testers.
Hi
I've been using Beans PropBasic for ages using 'Geanie' and having LOADS of fun. I had a need to talk to a chip (CH375) that reads/writes to usb memory sticks using 9 bit serial async rather than the usual 8, and just looked at the assembler code generated by PropBasic for 8 bit serial and tweaked the code to do what I wanted. A good intro to assembler with some codes requiring the manual to figure out- all good fun!
Dave
-Phil
I'm certainly not bored but it seems a lot of talent has been expended and is still being expended on P2 even though it's still in a state of both flux and fog. P2 will certainly get my attention and energy when I can see that I can actually build stuff with it, until then I need to focus on getting the most out of P1 because that's what I build stuff with.
@Dave Hein: 850,000 combined views for P2 clearly demonstrates how much the P2 hype has taken away from the current and reigning champion which is a shame because P2 while more than a twinkle in its father's eye, is still in it's rather prolonged gestation period and even after birthing will still be wet behind the ears.
While I have not tried your Tachyon, I do look at your thread from time to time to see how much you have achieved, all on your own!
There are many other interesting things to do, and I am sure that has taken some away from the P1. While I would like the P2 for some projects (yes, ZiCog & CPM is in there too), the P1 still has the same appeal as ever. Unfortunately things that the P1 can do well (especially 3D printer and Quadcopter) I just have not had the time to explore. I even have pcbs done
I really wanted to continue with my Prop OS version, but lack of time, and lack of interest, keep this project on the backburner. I wanted to add Michael's Sphinx compiler to it too.
Phil's scope project was an inspiration. I did some work to add-on to Phil's work but interest declined when Phil was side tracked with other projects. Hopefully Phil will resurrect this project at some time in the future.
Most of those drooling over the P2 are probably not the mass users Parallax need for its success. They are likely the silent ones, mostly absent from these forums. It's a shame we don't have a simpler spec P2 now, and the real P2/P3 can come a little later. Chip and Ken understand the minimum specs for a simpler P2 that their customers wanted yesterday.
Phil is both a VIP and a VIC in that he is a very important contributor in terms of quantity and quality of projects. Not just that but they are always very interesting too. However for the "P2 droolers" perhaps if they lent their talents in a balanced way we could keep the home fires burning brightly while cheerfully warming ourselves around it rather being forced to shiver in the gradually cooling winds while fanning the remaining embers. Surely the more that are still around and active on P1 when P2 comes out, then greater the exposure and success of P2 itself. Forumistas, don't neglect the here and now by being too focused on the dream waiting for it to come true. When it does, it does, and it will be a dream come true.
BTW, why haven't you tried Tachyon yet? You've got no excuse since there are binaries sitting there waiting that will run on the most basic systems.
The Raspberry Pi came out. Well there was a shiny thing if ever I saw one. Then the new improved Pi B+. Now the even more awesome Pi 2. I like to keep the Prop 1 in the picture when mentioning those as I think the combination of Pi for the heavy weight Linux luxury and the Prop as a real-time interface adapter for it is a great idea. We spend a lot of time getting the Prop tools running on the Pi. If we could sell that idea to just 10% of Raspi users that would be 500,000 Propellers sold!
The Espruino came out. Ooo another shiny thing. Well, the Espruino is a wonderful and inspirational thing. Far away from your cup of tea Peter but certainly the easiest way to get people into programming gadgets I have ever seen. I can't really keep that on topic Propeller wise, except for the tantalizing prospect of one day having the Espruino JavaScript engine running on the P2.
The Micro Python came out. OK, I did not get into that one. That is a P language and besides I had already backed Espruino on Kickstarter.
The Parallella came out. Hmm..well I backed that as well. Turns out the thing is still in it's box here as I have never found the time or enthusiasm to dig into it.
Just yesterday I found out that that it is quite likely that by the end of the year a Raspberry Pi like board could be available using a new SoC based on the RISC V architecture. I guess nobody has heard of RISC V yet but it is likely to be huge in the not so far future. The plans for that particular RISC V chip include, wait for it, multiple independent, real time, peripheral processors, with deterministic timing. Software defined peripherals, like the Prop and XMOS. That would make the idea of a P2 as a real-time interface to a Linux running SoC redundant.
That paper describes both the multiple cores, which look similar in function to what a Propeller connected to a CPU might do, and the idea of flagged memory capable of initiating an exception on read, write, or both per address.
RISC V and the LowRISC started to look like interesting possibilities that are likely to become reality, rather than just research curiosities when I discovered two things:
1) RISC V is a huge project going on at Berkeley and elsewhere with support from DARPA but more importantly, huge companies like Google, MS, and others. Those guys are seriously looking at an open instruction set design to replace Intel's expensive power guzzlers in data centres. Then there is support from those wanting to build mobile phones and IoT devices that need to be small and low power, oh and those guys want to get away from licensing ARM. Then there is the government of India, they want an open instruction set for security reasons, they don't trust Intel chips not to have back doors. Strangely, having said all that, Intel is also a backer, I guess they want to be in the game when the world flips away from x86.
2) Alex Bradbury, co-author of that paper and another guy on the LowRISC project are heavily involved with the Rasperry Pi.
In this video about LowRISC https://www.youtube.com/watch?v=r1i9SAOdyS4 he describes their planned LowRISC board as a "Raspberry Pi for grown ups".
That leads me to speculate that there may well be a new Raspberry Pi from the Raspberry Pi foundation itself that is based on LowRISC.
RISC V is clever in that it's base instruction set is only 40 instructions. Enough to run compilers, tools, Linux. After that it's extensible to 64 and even 128 bits! And extensible for floating point etc.
Perhaps Chip should switch the COGs to the RISC V instruction set for the P2
If you are interested in RISC V and LowRISC you might like to watch these videos from the first RISC V conference: https://www.youtube.com/channel/UC5gLmcFuvdGbajs4VL-WU3g Seems they have all the compilers, libraries, Linux and simulators running already. They have RISC V in FPGA. They have a new hardware design language for it all. It's amazing what they have already.
I had not seen the first conference. Added that to my list to watch soon.
Yeah, Intel knows it's a long shot, but if an open ISA were to catch on, they still have very significant fab and engineering advantages. It's smart to put a toe in, just to keep up with the goings on.
Firstly the idea that for most applications the Instruction Set Architecture does not matter. CISC vs RISC no matter. There so many other influences on performance and energy consumption. From the algorithms you use, to the code you write, to the compiler optimizations to the ISA implementation to the chip process technology. So why not have a common standard ISA, and make it as small and simple as possible for security and validation purposes? And also to get rid of the cost of endlessly creating new tools for new ISAs as we have done for decades.
Secondly that the ARM ISA is massive. They quote 3000 pages of instruction set description. There is no way ARMs famed low power consumption is down to it being a simple RISC machine.
Thirdly that the concept that the instruction set is the most important interface in the whole of computing. It's where all software meets all hardware! We have standards for so many other interfaces in technology from the old S100 bus to mobile phone protocols. So how come we don't have an open standard instruction set?!
Those Intel/ARM gravity wells are deep. But then so are the pockets of the big data centre providers. Google, Amazon, MS, FaceBook, others. They have realized the huge benefits to be had by using Free and Open Source software for years. So much so that they cooperate with each other on developing such free/Free software. It could well be time that they turn their attention to that hardware problem they have. With that realization of the power of cooperation they could well make RISC V huge in a short time.
Meanwhile. Other chip vendors are also huge and looking to find a market in tiny cheap IoT devices. They also understand the cooperative advantage of open source platforms, embedded Linux is everywhere for example. To get cheap and quick development they need to free themselves of the ARM license problem. They may well also be ready to pitch in together and make RISC V huge very quickly.
Interesting times.
Intel still produce 8051's, buried in their chipsets, so they still understand deeply embedded engineering.
That is a really interesting point. I've been thinking what it would take to run that touchscreen and operating system with two propellers rather than one propeller and lots of support chips. The hardware ought to be pretty simple - just two props communicating with (probably) 2 wires and then the automatic download from one prop to the other which has been around since the propeller was born.
The harder part is software but it ought to be fairly straightforward - what you would need to do is take an IDE like the propeller IDE, and each tab has an object in it and you have all the objects open as you write code and flip between them, but then you add a few lines when an object is loaded telling the IDE which propeller to load it to. Then the download process is still just one keypress. So then you can play around with 16 cogs running in parallel and you can experiment with writing the code for the Rx and Tx part of the comms within one IDE and debugging would get a lot easier.
If debugging is easy, then others would find it simple to write different objects like you say, and you could for instance have an object that uses an arbitrary number of communication pins where more pins = faster speed.
It is the debugging that is hard when thinking about code for both propellers at the same time, especially if you have to keep opening and closing files all the time.
It would be cool to have a virtual 64 pin propeller with 64k of ram. I think it is just tweaks to the IDE.
1) Whatcha flyin'?
2) Can ya do that sober, or do ya have to get ... ?
3) So, where'dja get yer "stuff?"
4) How much didja buy?
5) Can ya tell me where ya got it?
Am I close?
-Phil
Let me tell you about the thought I had when I wrote that comment.....
Let's say that you have two props side by side, inter-connectected by 8 IO pins on each prop, with one being a master and the other a slave. The slave would simply provide input to the master, using OUTA on the 8 pins. The master would have a cog running a continous loop to perform a range read of INA on the 8 pins connected to the slave. The master could then perform a wide variety of functions and the function performed would be based upon the result of the range read.
I am not sure whether there would be any benefit to this, as compared to just sending a value serially over one pin, but it was a thought, and I wondered if it had any merit.
EDIT: Providing that the idea has merit, an additional 2 IO pins could be used between the master and slave for full duplex serial communication, for other communication requirements.
EDIT: Now let's say (disregarding FDS and those two pins) that you have 9 inter-connected pins from the slave to the master, with the ninth pin specifically designated as a sign indicator, plus or minus. This ninth pin would nearly double the possible number of functions that could be performed by the master.
EDIT: Of course, you could just use 4 pins to set the inputs on the master, with another pin as a sign indicator, but then the possible number of results drastically decrease. But you all knew that already
These inter-prop communication systems have been discussed here many time over the years.
Might want an 8 bit bus between Props with two hand shake/clock signals. That's 10 pins on each Prop. We started out with 65 pins (I like to keep programming and EPROM pins free) and now we have 45 pins free.
To my mind this is grossly waste full of pins.
I'd go for a simple full duplex serial link unless speed is a major requirement. Still, one could probably do a half-duplex serial link using only one pin on each Prop that would get you a good turn of speed. Or or perhaps have a synchronous half duplex scheme with a data and a clock pin for even more speed.
All depends what you want to do.
Tackyon's top position is due to Peter J being so clever. He has channeled all documention and support through the one thread -- which I expect may be the longest ever.
Happy to see Dave Hein point out that the P2 threads account for something more.