MicroPython Forum (Archive)

EDIT: This is intended to be a general MicroPython publicity topic.

==========================

I've been doing regular "media research" for MicroPython on google, and while there's number, slowy growing, of mentions in embedded-related media, I never really saw any discussion in Python-focused mailing lists or blogs. Today I finally explicitly searched on https://mail.python.org/pipermail/python-list/ - and nope, no relevant hits at all.

Ok, so what's the plan here? Python people are gotta be ones interested in MicroPython, and same "delayed involvement" rules-of-thumb apply: you announce project now, in half-year people start to seriously look into it. Of course, there's risk that announcement will lead to immediate heated discussion which needs to be kept up .

So, this is question to Damien, who has a prerogative to announce the project, but given that he's very busy with technicals of project, I wonder if there're people who are on python-list and know MicroPython well enough to start/keep up discussion there? (Just in case, I am not - given resources to write a line, that's gotta be the line of MicroPython code, so I don't have free resouces for much of anything else.)

I'm going to be at PyCon in Montreal in mid-April, and I'll definitely be showing off my custom MicroPython boards, and my robot uCee that uses them. Hopefully a lot of people there will spread the word.

I just got a new order of boards back from SeeedStudio (revision 2) for my board:



I'll have a couple of those populated (one in the robot), and will be able to show the REPL and some simple control stuff with it.

- Jon

Before we start pushing Micro Python further there are 2 basic hurdles that I want to clear: 1) deliver fully on the Kickstarter, 2) make uPy run a decent amount of existing Python code. The former is easy to measure when done, the latter, not.

I was in contact briefly with Guido van Rossum months ago, and he knows about the project, which is a good start. I had an invite to the Django conference in the UK in Feb, to talk about uPy, but had to decline. I think I have an invite somewhere in my inbox for PyCon UK this September, which I'll try to attend. There was some initial negative response from a Python forum/user group during the KS, but I can't recall the details of that.

I don't have familiarity with the python-list, but I can eventually announce something on there.

If Micro Python could run Django, even in some heavily stripped down form, then that would get people's attention. Imagine nginx + uPy

Thanks to pfalcon, Micro Python can already act as a simple web server or client. With generator support, it can almost start to be a real server.

There was some initial negative response from a Python forum/user group during the KS, but I can't recall the details of that.

I guess that's quite expectable, with 2 obvious questions to be asked: "Why it's needed if there's CPython?" and "Why not revive PyMite?". So, in anticipation of that, announcement may emphasize key uPy differences. And otherwise, any controversy in mailing list discussion is good - the more flame, the more people will see it .

Before we start pushing Micro Python further there are 2 basic hurdles that I want to clear: 1) deliver fully on the Kickstarter, 2) make uPy run a decent amount of existing Python code.

I cannot comment on 1), but I can't think why it would be a blocker if source code was made open-source awhile. And as for 2), I understand, that it's not a key point (lack of resource to do everything at the same time is), but otherwise, it's not good approach for open source project . Being able to run existing apps is kind of an end (hopefully, just of some milestone), not the beginning. And it make sense to invite people to take a ride along, not show them photos post-factum. Anyway, my idea is to find more hackers for uPy, not just users. And if 1 in a 1000 of who looked at it will hack on it, then millions should see it .

So, let me put it like that: if you'd prefer 3rd parties not make any noise about uPy before Kickstarter delivery, please let us know here. Otherwise, even I may give up and post something somewhere . But if you somehow managed to refrain announcing it on Python lists in the support of KS campaign, then there should not be any reason to be shy with that still .

If Micro Python could run Django

Well, somehow I don't find that combo too exciting . uPy + Bottle or uPy + Flask - another matter. Anyway I indeed was interested in using uPy for web microframeworking since the beginning, regularly scaling landscape searching both existing frameworks and looking for new, etc. I still didn't find something really suitable for uPy (like, scaling down to MCU usage), and set to put together something myself, but making existing microframework run on uPy seems like good way to realworld-proof it.

I mentioned Django simply because it's got a lot of existing users, and if uPy can help them run their websites a bit faster with less memory, that would be a good place to start promoting uPy to the wider audience.

if you'd prefer 3rd parties not make any noise about uPy before Kickstarter delivery, please let us know here

I would like to have some control over how uPy evolves post-KS, and I don't have time to do that while KS is still on my mind. I don't want to be inundated with all mater of things and not be able to properly deal with them.

I need a few weeks, then you can go all out announcing

Sounds good, I'm glad I brought up this discussion, to avoid awkward situations.

Damien announced MicroPython on python-list earlier this week: https://mail.python.org/pipermail/pytho ... 72994.html . Discussion quickly moved from MicroPython itself to discussion of Python3 vs Unicode, which is common flame topic there nowadays.

Related thread was also started (by interested community member(s), which is good sign) on python-dev: https://mail.python.org/pipermail/pytho ... 34718.html . Again, it went as far as questioning why did Python3 did all that unicode to its loyal users .

Shortly after the announcement, it went on HackerNews: https://news.ycombinator.com/item?id=7840566 and stayed good deal of the day at the top. It's not the first time uPy is on HN (I believe original Kickstarter and initial source code release were there too), and there's consistently growing interest (440 pints/93 comments this time). I also saw some of contributors on the comments - thanks (I then concentrated on Python list replies).

We were also in Github "trending projects" list, reaching #2. We also surpassed 1K stars, 200 watches, 125 forks there.

Overall, I would say the responsive is productively good - there were few undue optimistic expectations expressed, but overall, it's more of technical interest and good community relation. I *absolutely* didn't see anything which could be classified as negative response, despite the fact that I expect responses along the lines of "why the heck it's needed", "why the heck you don't work with older projects", "Python on MCU is ridiculous", "why the heck it's needed if there's Lua", etc. Again, nothing like that at all. Only technical or just neutral comparisons and affirmation that it's good idea for MicroPython to have its own implementation way.

With all surface value above, we of course looked to grow our community, in particular, developers and contributors. Some results are seen already - there're various cleanups and portability patches, and interested community members started on adding Unicode support: https://github.com/micropython/micropython/issues/657 , https://github.com/micropython/micropython/pull/671


So, we're on good way to make the project sustainable, and let's follow this way - please post about your MicroPython-related projects on your blog and on forum: http://forum.micropython.org/viewforum.php?f=5 , mention it on other forums, etc.

Thanks to everyone involved!