[Mizar32-riot] Riot-avr32 - Interpreters

[Sergio Sorrenti]

Hi George,

yes, is correct, after the port,
the interpreters, but we have to
study how Zephyr does, or find clever ways
if we follow zephyr we have to focus on mutithread
interpreters where possible to not bloat the memory.

For us 1 atom of code is 1 thread...

Riot that is thread based(not process based),
also offer the minimum expense of memory per thread
(<25 byte per thread) among all RTOS.

I've think thatsomehow that is a it stupid
fix the number of supported languages

So we start and we do not know where we will finish,
also seeking in external support 1 person could join
in future to support a language.

What i want to target is a firmware for the basae OS,
then the full automation written open source or free
software no compilation at all. So in the structure, so
we use interpreters.

Isn't fast but will be fast... in near future.

I've think that the first to support could be
Micropython, then Lisp, then Javascript

we also have to not forget Open-PLC

in order to fill the gap with old automation.

Then nothing is fixed. Haskell could be too.

The functional programming is the future,

ther is also F# that is not bad, and is advancing in popularity.

About this thing of the Multithreaded enabled interpreter what you think?

Seem is the best way to go,
1 thread in interpreter become 1 thread in RTOS
as Zephyr does.

How to do for example with non mutithread interpreters?

For Lua we have PThreads a patch that make it multithread.

There is something like this in plan for picolisp or others?

Another important Interpreter could be Forth, that assure
the speed of compiled C, but is still running as source

So at the time of writing, Micropython, Lua(with patch), NewLisp
are Multithread.

So te simple way is wait other interpreters become so,
since the effort to make them multithread i do not know
how much complex it is or in such sense we can not do all
but do surgically what needed.

So sometime wait outside world maturation is important.

Is not important in 4 years we have 6 interpreters running...

Maybe only 3 is just fine. While wait the outside world

maturate and those interpreters are made multithread.

But you may have other clever solution looking at Riot
what can do what can not do , my view of non-programmer
is limited.

So i pass the hands to your programmers view

What you think?

Sergio

George Orais <gporais at yahoo.com> ha scritto:

> Hi Sergio,
> Thank you for sharing these info, very interesting and all noted.
> Btw want to confirm if my understanding is correct, first goal is  
> port Riot to Mizar32.After this we will port existing interpreters  
> (Lua, C, PicoLisp, TinyScheme) from Alcor6L to run on top Riot,  
> correct?After this, add Javascript, Python and Haskell interpreters  
> on top Riot, correct?
>
>
> Hi Marcus,
> Thanks for all these details about Riot port for AVR32, all noted.
> Let me dig into this and prepare my environment based on these  
> details, will get back as soon as I got questions or if its in ready  
> state.
> For the naming, I agree with the points on Riot- first or Mizar32-  
> first, but if you and Sergio votes for Mizar32-riot, I'm fine that ;)
>
>
> BR,Geo
>     On Tuesday, 17 November 2020, 04:20:40 am GMT+9, Sergio Sorrenti  
> <sergio at simplemachines.it> wrote:
>
>  Hi Marcus and All,
>
> regarding Riot OS i planned to make a fork when
>
> we start to have an RTOS + Interpreters.

-- 
Sergio Sorrenti - SS - Associazione Simplemachines - President
e-Mails: (Dev): sergio at simplemachines.it | (Bizn) kansei.studio
Personal: sergio at onipv4.org | GPG-Signed: questo.mondo at libero.it
GSM-Telephone, Telegram , Whatsapp , Skype (Web): +39 3492560300
My Projects: | GNU Soul | § SgrA* | 🕎 Nwiw | ツ Share the time