Firing up my RP2040
Brought to you by:
m-atthias
Menu
▾
▴
I strongly encourage diversity in Forth, as different implementations have different goals and special features that should be explored further. My own implementations have a strong focus on compiler design, whereas Tachyon has a very innovative way to handle the area where the programmer directly touches the hardware, with great ideas in the user interface design. Amforth, on the other hand, gave us recognisers and advanced the state of the art in parsing. I know from the Flashforth maillist that many users switched from there to Mecrisp-Stellaris instead, and I am sorry for that, because Mikael Nordmann has done a very good traditional Forth, and took great care in maintaining it for such a long time. There are many more, please do not consider this small list as complete.
So every great implementation adds its ideas and advancements. For example an implementation written or adapted for a particular chip in mind will integrate much better with the capabilities of the hardware than a Forth which was designed to use a common minimal set of peripheral support found everywhere. Mecrisp-Stellaris is of the "generic, run everywhere" type, whereas Tachyon on Propeller P2 exploits almost everything the underlying chip has to offer.
Probably the most fulfilling acknowledgement is seeing other Forths adopt ones ideas - see Gforth experimenting with constant folding these days, for example, and recognizers as part of the Forth standard. Or Mecrisp-Stellaris choosen as base to launch Tachyon into a new architecture. Its OK then when individual implementations fade, as they served their purpose as a cornerstone in Forth development well, as did a FIG Forth assembler listing for me.
To summarise - keep being creative and inventive, and explore paths never been taken before.
PS: Peter, I am sorry I cannot give this discussion forum more of liveliness - I am already scratching at the top of what is possible in my time available besides tending my garden, my rabbits and working full time as FPGA developer for a space mission.
Thanks for that feedback Matthias, and a nice summary about diversity in Forth. I have had many criticisms that Tachyon is not ANS, yet if it were it might have faded like the Forths that were so.
I'd like to be able to build in your assembly mode compiler into Tachyon and just use it where I might have been tempted to convert a function to assembly code, but instead invoke the "Mecrisp" compiler to turn the source into assembly for me instead. Also I'd like to use the constant folding in a similar way with threaded code where I could perhaps use braces or something around it to indicate that the compiler should resolve it.
But I'm very appreciative of your hard work and making Mecrisp available for so many platforms, but you shouldn't be responsible for maintaining the "liveliness" of the forum. I would have thought with so many people using Mecrisp that they would give a bit back too. Even just some feedback of some kind.
Last edit: Peter Jakacki 2021-09-21
Actually, the active user base might be far less than one might think, unfortunately. There are many "downloaders", yes, but to transform from the "downloader" to the active user takes time and determination and many people just don't get that far quick enough and then abandon the whole thing.
I spent a few hours yesterday playing with your Tachyon just to make some simple blinking led patterns to blink exactly the way I wanted them to blink and had lots of fun with that and head scratching too. Now, if it wasn't Tachyon or Mecrisp or any other Forth for that matter, I could most probably google out the code in less than 30 minutes and "save" myself the trouble and then generate plenty of forum activity asking why the code doesn't work the way I wanted it to work :).
Posted Are you a Forth "user"? to Forth2020
I am not. I will never be a registered FB user and this group is a private one so can't even see the posts. I am yet unsure of what exactly the focus of this FB group is. I can see on this page it's mainly PC or esp32. I prefer open, public forums. Seem more sensible to me.
I'm sure it seemed a reasonable choice of platform at the time - get free hosting infrastructure etc. I notice Generation X's FB share is declining but the Boomer's share is increasing rapidly maybe this is partly a symptom of this. I doubt the founders of the group foresaw FB as being a contentious issue, making it a closed group probably seemed prudent - so no blame there.
However if they want younger blood in the Forth gene pool, FB could become a barrier and an attempt to '2020-ise' Forth might not be as effective as they wished. You can still watch their video conferences on Youtube and there is definately some good content - an interview with Leo Brodie stands out and some interesting demonstrations, but to actively participate you still need to swallow the distasteful FB pill. There is a push to get the ESP32 working with Forth but as I said above there is closed source Expressif code in the way of freedom.
Last edit: gustinian 2021-09-22
I'm not a FB user either, and never will be. I have my site https://mecrisp-stellaris-folkdoc.sourceforge.io/index.html and my IRC channel #forth-hardware-projects on irc.hackint.org.
My super sneaky all Mecrisp-Stellaris Forth project has been downloaded over 600 times by Bluepill users wanting to find out which MCU they actually have in their board.
https://sourceforge.net/projects/mecrisp-stellaris-folkdoc/files/bluepill-diagnostics-v1.6.zip
Personally I'll take four doers over 100 talkers.
Cheers,
Terry
As I have said many times before, I am not a fan of FB but I'm there because Forth2020 is there. I'm not an anti-FB or anti anything really although I choose what to follow or not, and my thinking is that we can whinge and whine about all manner of things, but all that happens is that we end up missing out. I have a friend who does not "believe" in cell phones for instance, whatever that means. Enough of that now.
Here's the original post:
Are you a Forth "user"?
Just been having a discussion with Matthias Koch and others on Mecrisp forums and it got me to thinking. I wonder just how many users there are of Mecrisp and Tachyon etc. I know a lot of people download these and "have a play" but mightn't go beyond that. It's a shame because these are powerful dev tools and at the very least are second to none when it comes to testing hardware.
I don't think that the learning curve is too steep so that the casual evaluator will abandon the whole thing and save themselves trouble. I've seen the learning curve working with IDEs and libraries and it is a hard road, even just to set them up properly.
Just thinking about what is lacking I realize that with these "popular" tools that many people will share code and publish how-to's and reviews. Sometimes the stuff they share is very simple and basic, but it is feedback and helps to make it popular. Because of that many others are willing to put in more than the usual effort to learn and use it. However with Forth although I have written and published many projects in the hope that it would encourage others to publish their project or at least tell a story about it, the sum total of all that feedback is effectively nada, nil, zero, squat.
So I encourage anyone who is enjoying "using" Forth to give back and talk about it, and maybe even submit it to Hackaday or a magazine or something. That would be great exposure. But remember, Think Forth (don't translate code) and keep it clean and simple by factoring code with minimum stack juggling and using short succinct names and methods. That is avoid_using_long_names like on_left_motor and off_right_motor etc when instead you can simply define on/off/left/right and say "ON LEFT MOTOR" or "OFF RIGHT MOTOR" etc. Forth is a language for creating an application specific language, so use it (and give some back).
All good points., Peter.
Here is some of what I think.
To get new people hooked to actually using Forth on a given micro we need:
1. a cheap, popular and easily obtainable hardware (both the "brain" board and the simple peripherals like LEDs, buttons, rotary encoder, motor, speaker etc.
2. a simple yet complete tutorial on how to get that Forth on that micro in a simple way (the process can't be complicated).
3. some very basic examples with simple step by step explanations what they do and how they work and how useful they can be in real life (road signs, disco lights, door/gate locks, garden sprinkler/shades control etc.)
4. some more advanced, thoroughly presented examples for the faster learning to show off how easily something more complicated can be accomplished with Forth
5. some real life project examples with board photos, youtube videos, stories behind these projects, anecdotes etc ( I really enjoyed watching your short YT videos or ZOOM presentation recordings)*
6. a public, easily accessible place/space/board for discussions about anything dedicated to embedded Forth (any embedded forth) and sharing the code there too.
7. a few dedicated people with enough time on their hands
Simple, cheap, useful, popular are key words here.
It seems like we already have most of the above (except maybe #3 & #4 & #6) but scattered among many different places.
And the new user, preferably with hair but without grey hair, already spoiled by Google, FB and YouTube, wants results cheap and fast and needs someone to talk to.
I can certainly offer my story on how I discovered Forth and how refreshing that experience has been and maybe a few simple examples as in #3.
Let's be realistic. Forth has little chance to be as much popular as C or Java or Python and even if it was to happen, it'll take years for sure. I don't like it but I accept it.
What we really miss are Forth success stories and I mean real stories about how using Forth has won the user something of monetary value. That could be a game changer if something really spectacular surfaced and people took notice.
BTW, in case you haven't noticed, we have a (just started) dedicated ForthSpace at Parallax forum but something more cross platform would be much needed.
Last edit: Maciek 2021-09-23
I see the RP2040 as an ideal replacement for the ageing P1 chip. It has over 8 times the RAM, and a stack of peripherals and the PIO state machines which together will do what we do with cogs. At 125MHz many of the instructions execute in 1 cycle which makes this a good 6 times faster than the P1 cog @80MHz/4 but much much faster than the P1 if it had to run code from hub.
While it doesn't replace the P2, it can do the job of a P2 with several of these chips for a fraction of the price and a fraction of the power consumption too with only a single 3.3V supply needed.
@Maciek - I haven't been on the Parallax forums since earlier in the year but Cluso fills me in at times and mentioned a Forth forum there. Funny thing is that at one time some wanted to setup a Forth sub-forum but I was dead against it because once you have your own private corner then Forth doesn't get anywhere near the exposure it would in a more general forum. However
Parallax are not supportive of Forth and avoid mentioning it despite the fact that it has so many capabilities built into it that really show off the P2.
Now, back to Mecrisp and I can't help getting sidetracked while I write my macroassembler. I just added ANSI controls and a highlighter to words as well as very easy BAUD settings. Just select UART0 or UART1 although UART0 is selected by default and set the baud rate.
Last edit: Peter Jakacki 2021-09-23
I'll be totally frank with you, Peter. It was your Tachyon and TAQOZ Forths that brought my attention to Parallax, not the other way around. Have there been no forth available for the P1 or P2 I wouldn't have even given them a second look as I do not like to get into the code-compile-flash-correct-repeat countless times cycle again. It was Mecrisp that brought me to msp430 micros and stm8eforth that led me to the stm8 chips. It was always forth that was a drive.
That said, do you really think Raspberry Pi Foundation (or Broadcom) will be any different ? I have no illusions. But there already is Mecrisp Stellaris and hopefully will be Tachyon available for the RP2040 so I might just as well go with that. My Digikey basket has it in for two weeks now and it's cheap so has a big potential for the first time forth experimenters. I do not particularly like the way ARM handles the IO's but it is what it is and it's usable.
That separate forth sub-forum thing - I disagree. It's much easier to find an interesting forth stuff in the forth sub-forum than in a general forum. And you can see people start posting in that forth sub-forum and participating and that is heartwarming and hopefully they spread the word and that's the most valuable and effective advertising (though very slow).
I'm not a missionary, I have no desire to save the world. I'll just carry on with forth and fun which is most of the same for me at this time.
One other thing to bring to your attention - I tried to post on your Tachyon page here on sourceforge numerous times but seems like I have no permissions to do so while I can freely create post here or reply to other posts. Have you banned me already ?
Last edit: Maciek 2021-09-23
Waaah! You can't post on my pages??? Don't tell me that's why nobody has posted there!
Yep, tried it on a private page and I had to sign in to post! Looks like it's fixed now.
Thanks for the heads-up.
Last edit: Peter Jakacki 2021-09-23
btw, the RPi foundation is not Broadcom although they have used Broadcom chips on the RPi "PC". Because they are a foundation they have invested in designing their own MCU chip of which the RP2040 is the first. The RPi Pico is simply their RP2040 board which I have, including a supply of chips. I don't think you can order these $1 chips from Digikey though but I have ordered them locally through official RPi distributors.
262KB RAM - Dual core M0+ @133MHz (single cycle) - single cycle I/O - USB OTG - 2 PIOs with 4 microprogrammed state machines each - etc
Last edit: Peter Jakacki 2021-09-23
Sure you can order them. I just looked. Cost is €0.85 for one, more for a whole reel of them. Mouser lists them for €1.10. Farnell wants 0,828 € in lots of ten.
The problem is that the number of chips in stock is precisely zero. Farnell says they expect some in November, but I can only assume they'll be sold out immediately.
I was talking about the Pi Pico and Maker Pi Pico, the exact same you showed in you video clip. These are plenty in stock, next week at my doorstep just to be able to follow your Tachyon endeavor. Bare chip means some effort to design and make a board for it which at this stage is not time/cost effective in singular quantity.
I know they're not Broadcom but I get the impression they've generated a lot of volume sales for them ah have close ties to say the least.
Sure you can order them but with "stock = 0 - immediate" that is a problem 😉
I have and can order them locally in small quantities which is sufficient for now.
TSMC manufactured the RP2040 in 40nm. Nothing to do with Broadcom.
Last edit: Peter Jakacki 2021-09-24
Peter, I keep thinking about your original FB post you reposted here titled "Are you a Forth "user"?" and can't help myself but to ask, what's your perception of the answers given there on FB to this post ? I'd like to think there were many but what do they say and what do you make out of it ?
As usual the main point is lost. As a question it wasn't really answered. Sure, people talk about what they do and what they think, but did they answer whether they are a "user" or a user. Not really. But then again it is not really about answering the question either, as this question is more rhetorical in nature and meant to stimulate users to action, to contribute in any little way they can so that they are not a "user" - that is - one who uses but doesn't contribute.
So as we saw with Mecrisp downloads, that they are there, but where is the feedback and the giveback????? How many users of Mecrisp are there really? Terry Porter gives back, some others do, but that is but a few.
One of the things I wanted to emphasize too is if we want to encourage others to understand Forth, then it is important that they think Forth, or more specifically not think in a box. Make Forth clear, clean, crisp (Mecrisp) and readable.
All right. Now I get it.
In my case I'm a bit shy to show my, often stupid or trivial, forth attempts . Then I'm just a hobbyist with not so much hobby time and making a good example to be shown publicly, even a simple one, without being ashamed of it means it not only needs to work but it needs to be perfect (if it is to be an example and not just any piece of code). How do I judge if it is perfect when I am still an inexperienced experimenter ?
Code will never ever be perfect but if you did have perfect code then no one else would ever show their less than perfect code. When you share your "hobbyist" code it encourages others especially those at the same level. There are many levels and some go sideways too. When you do share, it causes you to look at your code in a different light and you see things that can be improved and made clearer etc.
Hear! Hear!
Peter, I couldn't agree more and discovered I had some tips to contribute myself, far more than I can fit here so I wrote an article on my site at :
https://mecrisp-stellaris-folkdoc.sourceforge.io/tell-the-world-about-forth.html
Cheers,
Terry
That makes things a whole lot easier and it is encouraging too. Thank you.
I'll give it a shot. I'll do as best as I can and let's see what responses, if any, I get. Maybe that gets some discussion going even if only to laugh me out of court. I'll take that chance.
Last edit: Maciek 2021-09-29
And so here I am with my super tiny contribution effort to the Forth community. As this is a Tachyon Forth example and story, I have decided to post it at Peters' site here.
Nice Job Maciek,
I for one enjoyed reading your well written PDF.
Cheers,
Terry
Thanks, Terry.
Especially that it comes from you ! Many, many clicks on your sites are mine and I recall we had some email exchanges some (many in fact) months ago. A little pity the original Mecrisp for the MSP430 has remained untouched for years now but I can understand a reason for that. Very nice chips, these little msp430's but a bit expensive these days compared to the lower end ARM ones.