1. Forum
  2. Usenet
  3. COMP.LANG.FORTH

  • Silly question on tethered Forth

    From Jose Morales@21:1/5 to All on Wed Sep 7 07:32:13 2022
    Hi all, I’m new to Forth after rediscovering it on embedded systems.

    I have a quick idea vision I was wondering if any of the Forth builders are considering.

    Let me start with the vision, I found that one of the biggest value for Forth as part if interactivity with the the little systems was the fact that one can basically do this magic (I say that with my C hat on) where you can actively stop a loop,
    reprogram a “function” and restart your loop. Which is why most MCU’s get a FW upgrade, they are basically refactoring or adding functions to a loop of tasks or functions…

    OK, so here is the idea, how can one automate this in a way there you add a upgrade task to your loop to enable OTA upgrades… here I’m really considering, can a PRECOMPILED upgrade of those functions *or words* be sent to a FORTH system without
    interacting with the forth console?

    Like I said, I’m still learning Forth, but it would certainly just trigger an update to the dictionary, same as the ; does…

    Here os a problem (no not one I’m working on or considering but a really problem non the less) your electric meter, that is one of the most common IOT device out there, seen that those range in the billions, upgrading those with even large pipes must
    be a PITA. This type of system could benefit massively with these “simple” updates at scale. Even at say, 4MB Full RTOS + App, is a very large waste of bandwidth… vs say tens of kilobytes with that I’m envisioning.

    The other problem to that coin is, there are probably many targets types or versions of the MCU’s, so as a reference, consider you are developing in a tethered Forth environment where you are actively testing gains all your target types on a single
    code base. You can keep all your *includes* for each target separate, they are just the HAL, but your app is a single source.

    For me, in general, I’m working on STM32 *3 different flavors* PIC24 and AVR MEGA and Dx to that, I’m adding M0+ *RP2040 and RISC/V which will be like another multiple chip from an architecture perspective, just like the M0 and M0+ variances. I’d
    like to target these different MCU within a single code base. Sure I still have to maned the HAL drivers/Compilers, but it would be interesting.

    It looks like there is something similar to this tethered option, MikroForth looks like it has the basis of what this would be the Ideal IDE only if it supported all these targets.

    Thought, counter, or pointing me in project that is set out to this would be great.

    Cheers!

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Marcel Hendrix@21:1/5 to jem...@gmail.com on Wed Sep 7 09:32:51 2022
    On Wednesday, September 7, 2022 at 4:32:15 PM UTC+2, jem...@gmail.com wrote: [..]
    you can actively stop a loop, reprogram a “function” and restart your loop.
    [..]
    I assume you have a main word A with that big loop inside of it. Inside this loop
    a word named B is called. Something stops A, writes / uploads a new definition named B, and executes A to restart the loop.

    No, this will not work as if the new B replaces the old B, unless run on a Jupiter Ace. However, when B is a deferred word, then ": new_b ... ; IS B A" would work on any Forth without dirty tricks. (I'm sure somebody is
    going to teach you that dirty trick now).

    -marcel

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Paul Rubin@21:1/5 to Jose Morales on Wed Sep 7 13:09:52 2022
    Jose Morales <jemo07@gmail.com> writes:
    can a PRECOMPILED upgrade of those functions *or words* be sent to a
    FORTH system without interacting with the forth console?

    There aren't any built-in Forth features for this, so you'd have to
    engineer something for your specific environment and application, but
    Forth is low-level enough that all things are possible. Simply
    replacing a word in the Forth dictionary won't work unless the word is DEFERred, because otherwise even if you replace the word, all previously existing references will still point to the old version.

    What you really want is Erlang (erlang.org). It is designed to do
    exactly what you are asking and has well developed features for the
    purpose. It won't run on those little MCU's (it needs some megabytes of
    ram), but you could still do pretty well by studying it and playing with
    it on a bigger computer. I think that an implementation on the ESP32 or
    RP2040 class of MCU's is feasible, but the current main implementation
    is intended for larger machines.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From dxforth@21:1/5 to Marcel Hendrix on Thu Sep 8 10:45:55 2022
    On 8/09/2022 2:32 am, Marcel Hendrix wrote:
    On Wednesday, September 7, 2022 at 4:32:15 PM UTC+2, jem...@gmail.com wrote: [..]
    you can actively stop a loop, reprogram a “function” and restart your loop.
    [..]
    I assume you have a main word A with that big loop inside of it. Inside this loop
    a word named B is called. Something stops A, writes / uploads a new definition
    named B, and executes A to restart the loop.

    No, this will not work as if the new B replaces the old B, unless run on a Jupiter Ace. However, when B is a deferred word, then ": new_b ... ; IS B A" would work on any Forth without dirty tricks. (I'm sure somebody is
    going to teach you that dirty trick now).

    "main word A" in the previous paragraph sounds a lot like the forth interpreter :)

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From =?UTF-8?B?4oCi?=@21:1/5 to jem...@gmail.com on Sun Sep 11 22:32:21 2022
    On Thursday, 8 September 2022 at 12:32:15 am UTC+10, jem...@gmail.com wrote:
    Hi all, I’m new to Forth after rediscovering it on embedded systems.

    I have a quick idea vision I was wondering if any of the Forth builders are considering.

    Let me start with the vision, I found that one of the biggest value for Forth as part if interactivity with the the little systems was the fact that one can basically do this magic (I say that with my C hat on) where you can actively stop a loop,
    reprogram a “function” and restart your loop. Which is why most MCU’s get a FW upgrade, they are basically refactoring or adding functions to a loop of tasks or functions…

    OK, so here is the idea, how can one automate this in a way there you add a upgrade task to your loop to enable OTA upgrades… here I’m really considering, can a PRECOMPILED upgrade of those functions *or words* be sent to a FORTH system without
    interacting with the forth console?

    Like I said, I’m still learning Forth, but it would certainly just trigger an update to the dictionary, same as the ; does…

    Here os a problem (no not one I’m working on or considering but a really problem non the less) your electric meter, that is one of the most common IOT device out there, seen that those range in the billions, upgrading those with even large pipes must
    be a PITA. This type of system could benefit massively with these “simple” updates at scale. Even at say, 4MB Full RTOS + App, is a very large waste of bandwidth… vs say tens of kilobytes with that I’m envisioning.

    The other problem to that coin is, there are probably many targets types or versions of the MCU’s, so as a reference, consider you are developing in a tethered Forth environment where you are actively testing gains all your target types on a single
    code base. You can keep all your *includes* for each target separate, they are just the HAL, but your app is a single source.

    For me, in general, I’m working on STM32 *3 different flavors* PIC24 and AVR MEGA and Dx to that, I’m adding M0+ *RP2040 and RISC/V which will be like another multiple chip from an architecture perspective, just like the M0 and M0+ variances. I’d
    like to target these different MCU within a single code base. Sure I still have to maned the HAL drivers/Compilers, but it would be interesting.

    It looks like there is something similar to this tethered option, MikroForth looks like it has the basis of what this would be the Ideal IDE only if it supported all these targets.

    Thought, counter, or pointing me in project that is set out to this would be great.

    Cheers!

    Looks good.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)