Sad farewell to pyboard and micropython, another push is needed to finish things off...

General discussions and questions abound development of code with MicroPython that is not hardware specific.
Target audience: MicroPython Users.
adouglas
Posts: 16
Joined: Fri Apr 30, 2021 4:46 pm

Sad farewell to pyboard and micropython, another push is needed to finish things off...

Post by adouglas » Mon May 03, 2021 11:35 pm

I think it would be a good idea to do a gofundme or something to fund another round of development of the pyboard. I had to give up on the pyboard and micropython, because of the insufficient documentation; it takes so long to try to learn how to do things, and how the system works, that the advantages of python become lost.

Also to mop up the last of the bugs... there are lots of features, but even some of the demos in the tutorials don't work, and there are a fraction as many tutorials as are needed.

I tried to keep track of things, but I got such a laundry list of complaints I decided I could not justify pouring my time into learning this platform, it will take too long. Even when I have poured all that time in, I would be on an island with a relatively small community and limited third party support, compared with arduino.

To solve the lack of community starts with improving documentation and streamlining the technology. The technology is impressive, but just not quite there yet, which is frustrating to no end, because I see this situation so often in community projects and open source stuff. We like to start projects and get them mostly working, but that few miles to get to a solid, reliable, durable, dependable system is rarely traversed.

If the system and documentation can just be finished up to a good state, then community will start to flow in, use will start to increase exponentially. When I browse tutorials on adafruit and elsewhere, there is a disproportionate number trying to use micropython; but they are all hamstrung. People buy the boards and do a little project, and that's it. They do not seem to do anything more. I assume this is because they encounter the same barriers I did.

Arduino's official documentation is also lacking in some ways, failing to discuss how some things work, everything from which pins support interrupts to whether a function is even blocking. But at least if I overcome those barriers by forming my own personal documentation, cobbled together from across the web, I feel like I will be in a better place, whereas with micropython, I am kind of stranded no matter how much I pour into cobbling together my own personal documentation hoard.

I know it's common to leave the user to go hunt down the details themselves, but I think it is in opposition to the batteries included mindset which I so like about python.

I think it has been too long with too little progress, I don't know what has been done since the kickstarter in 2016, but that was a long, long time ago. I assume a lot of bugs have been flattened and features added, but documentation is very basic and essential to deliver the value that python has for rapid learning and development.

stijn
Posts: 653
Joined: Thu Apr 24, 2014 9:13 am

Re: Sad farewell to pyboard and micropython, another push is needed to finish things off...

Post by stijn » Tue May 04, 2021 8:01 am

adouglas wrote:
Mon May 03, 2021 11:35 pm
People buy the boards and do a little project, and that's it. They do not seem to do anything more. I assume this is because they encounter the same barriers I did.
This gets stated as a fact, but without any data to back it up. Don't get me wrong: I'm sure this happens, but the real question is: how much does this happen compared to all people buying a board (not even specifically pyboard, just anything which runs MicroPython)? Pretty much impossible to answer, but don't forget we all sort of live in a bubble which makes it hard to form accurate ideas about a certain phenomenon if you don't try to look outside of that bubble.
I don't know what has been done since the kickstarter in 2016, but that was a long, long time ago. I assume a lot of bugs have been flattened and features added, but documentation is very basic and essential to deliver the value that python has for rapid learning and development.
Indeed a lot happened since 2016. Would be interesting to know what bits in the documentation exactly are lacking for your particular needs. There's no denying the documentation is lacking, but it's also not exactly such that the whole ecosystem is nearly impossible to work with (which is what one might think after reading your post), as proved by others being able to do things with it. Even just starting from scratch and by only using the existing documentation.

doublevee
Posts: 71
Joined: Mon Jul 02, 2018 11:09 pm

Re: Sad farewell to pyboard and micropython, another push is needed to finish things off...

Post by doublevee » Tue May 04, 2021 8:29 pm

As a very inexperienced code writer I have been teaching myself Python and learning MP as it is the perfect tool for someone such as myself to produce results that otherwise I couldn’t dream of.

I think anything embedded is going to be more complex than standard Python and there is unfortunately low level learning required for each microcontroller - this is inevitable.

Whilst I can sympathise with your comments regarding documentation, I don’t think you will find anyone on here that says it’s great. What I can say though is that it has got better and issues are picked up quickly and fixed/improved.

I feel your comments are slightly unfair and overly negative but I can totally align with your feeling of frustration regarding bugs and features. But consider the development speed and feature set (Asyncio alone is incredible) and the overall development is still very fast and productive.

The forum support is superb and I would urge you to stick with it a bit longer and reap the benefits.

User avatar
marfis
Posts: 206
Joined: Fri Oct 31, 2014 10:29 am
Location: Zurich / Switzerland

Re: Sad farewell to pyboard and micropython, another push is needed to finish things off...

Post by marfis » Tue May 04, 2021 9:12 pm

> People buy the boards and do a little project, and that's it. They do not seem to do anything more.

I felt this was a unfair statement so why not explaining how our experience is?

We‘re working in teams of 3..6 people developing FW entirely in uPy for a device thats pretty complex, for a large german industrial company. Think about a pyboard d sf6 loaded with several 100k’s of precompiled mpy files and handling several uarts / i2c buses and USB in parallel. uPy has handled this staff gracefully and reliably.

The relatively soft transition / improvements in uPy from 2016-now has helped us tremendously to periodically update uPy while still heavily developing the application code. I respect damiens work as a reviewer a lot and am more than happy to wait a bit longer until things are merged, rather then having to hunt down bugs or broken code due to API changes etc.

Docs are sometimes out of date, but they do keep getting better, with a lot of hard work from contributors.

And the development from 2016 onwards has seen LOTS of goodies, say for example the new asyncio v3 module just to start with, not to mention support for completly new MCUs such as the rp2040.

User avatar
rcolistete
Posts: 350
Joined: Thu Dec 31, 2015 3:12 pm
Location: Brazil
Contact:

Re: Sad farewell to pyboard and micropython, another push is needed to finish things off...

Post by rcolistete » Wed May 05, 2021 2:36 pm

There are many companies using MicroPython in many variants (official, Pycom, M5Stack, OpenMV, etc) and boards. Some applications are in the field, running for months or years.

So it is up to you to read, search, learn, practice, discuss, report bugs, etc, about MicroPython.
My "MicroPython Samples". My "MicroPython Firmwares" with many options (double precision, ulab, etc).

Delebel
Posts: 41
Joined: Thu May 25, 2017 2:21 pm
Contact:

Re: Sad farewell to pyboard and micropython, another push is needed to finish things off...

Post by Delebel » Wed May 05, 2021 4:44 pm

I fully understand your frustration and I've been there done that. However the reality is that no development platform is without some issues and my hats goes off to those that help out on any forum. I'm a C programmer from way back and moving to MicroPython was a choice of a lesser evil. Traditional IDE for controllers of any manufacturer requires installing huge software packages just to write a very basic "hello world" or blinky program. In contrast once you get a Repl prompt on a given MicroPython board your already in the driver seat. The learning curve is still long but at least the immediacy of the interpreter is much less daunting then getting 100 errors on a first compile in an IDE in C. Bottom line Micropython is not perfect but I am dedicated to persevere as I've had my share of loading an IDE and 3 hours later not having a single successful compile. Regards...Denis a newbie in MicroPython

adouglas
Posts: 16
Joined: Fri Apr 30, 2021 4:46 pm

Re: Sad farewell to pyboard and micropython, another push is needed to finish things off...

Post by adouglas » Wed May 05, 2021 9:06 pm

I do greatly appreciate python and micropython and the pyboard, sorry if I sounded negative. I greatly respect the project and you guys. However I do maintain that there is a distance that needs to be traversed from "it's working and can be used if you put enough sacrifice in/have a great deal of existing knowledge" and "highly durable, reliable system that is really a shining, solid, well developed tool".

I am not aware of how much labor time has gone into the development of micropython since 2016, but in the kickstarter it is literally one very talented and awesome guy that spent six months, and got paid something like $70,000. The kickstarter is approx 95,000 uk pounds, so suppose convert the currency and subtract the costs of paying for the rewards the kickstarter contributors get.

It is extremely impressive how much got done with so little, but there is a certain imbalance with the time spent developing the tool and the time spent using it, if you guys are using it to do big serious projects with especially. To be clear, I think it deserves more development time.

I don't think being able to use it for big serious projects is neccesarilly a good way to benchmark the state of development. The amount of time it takes to become acquainted with all the details, bugs and all, dig up the information needed to do everything from clearing interrupts to identifying which objects can be used in which context, which functions are blocking and which not etc.use it might be within profitable limits, but it still seems to be much more than is it needs to be.

I really appreciate the promise the system has, but am often frustrated by technological projects from the community, whether they are facebook alternatives or microcontroller boards, because of this tendency to not traverse the last few miles to a solid, well worked out, refined by many labor hours and perspectives, system. There are counterexambles, like LinuxCNC, and of course many other open source software systems.

I would love to spend some months of my life being a part of such an initiative to further improve the documentation, software and maybe the hardware a bit (the hardware seems pretty good mind you). Not that I am a great person to do such a thing, but as an extra pair hands with a certain perspective I mean... That's why I post, my heart is still captured by the promise here. Maybe someday soon we can do a gofundme and get some people together to do another big push.

I believe the micropython and the pyboard and maybe esp8266 boards could really catch on like wildfire if we could traverse this last little distance. If I could invest in it, I would... it's ready to explode here, but I think it needs a push or it will only continue to putter along at the same rate for many years.

adouglas
Posts: 16
Joined: Fri Apr 30, 2021 4:46 pm

Re: Sad farewell to pyboard and micropython, another push is needed to finish things off...

Post by adouglas » Mon May 10, 2021 1:16 am

update: after exploring arduino, I have come back to the pyboard. The most rational decision is to use arduino, but I just don't feel like it's as fun. The pyboard will probably work well enough for my purposes and I would rather learn python. I might switch back in a month, depends how it goes.

karfas
Posts: 33
Joined: Sat Jan 16, 2021 12:53 pm

Re: Sad farewell to pyboard and micropython, another push is needed to finish things off...

Post by karfas » Tue May 11, 2021 9:39 am

Maybe I'm on the way to become the grumpy old man complaining about today's young people.

Obviously it is YOUR responsibility to choose a development environment. If you don't like the way product X is managed/documented/coded you are free to use something else.
It is also YOUR responsibility to gather knowledge about the hardware and all software layers between your application and the hardware.

No currently existing development environment (be it Micropython, Arduino, mruby, AtomVM, whatever) or documentation will replace YOUR task to read and understand all information in reach - including hardware datasheets, API descriptions and the source code of the environment you had selected.

But this is time-consuming, hard work, isn't it ? Surprise!
No working code examples in the third google link...

Please realize that developing for embedded devices IS more difficult and requires way more knowledge, practice and tests than developing for largely standardized hardware and operating systems.

So, micropython is not documented enough, too buggy, too complicated ?
Yes, all this might be true.
However, as this is open source, YOU can change this - create issues, create patch requests, discuss in the forum, read the source.

Too much work/effort ? Maybe, but why do you expect from everybody else to make micropython a better experience for YOU ?

Please stop bashing micropython for the decisions YOU made. This is not constructive.
My repositories: https://github.com/karfas

SpotlightKid
Posts: 459
Joined: Wed Apr 08, 2015 5:19 am

Re: Sad farewell to pyboard and micropython, another push is needed to finish things off...

Post by SpotlightKid » Tue May 11, 2021 11:40 am

karfas wrote:
Tue May 11, 2021 9:39 am
Please stop bashing micropython for the decisions YOU made. This is not constructive.
This is not constructive either. MicroPython has many flaws and not discussing them doesn't help anybody. Nor does admonishing people who make the effort of explaining their reasons for not liking of abandoning MicroPython instead of just leaving. Nobody likes if somebody criticizes their favourite tool. But the mature thing is to listen and learn.

Post Reply