Main Site Documentation

Second Preview of TinyCLR OS Core Features


#1

Today we are excited to release the second preview of our TinyCLR OS. It works great with the community edition of the just-released Visual Studio 2017 from Microsoft. All you need to do to get going is install the latest extension, update your device’s firmware, and drop the new NuGet packages in your local feed. If you’re just now getting started with TinyCLR or just want a quick refresh, take a look at the first news post and the release notes for more information.

Some of the things that exist today in TinyCLR OS:
[ul]Stable preview firmware for G30, G80, G120, and products based on them.
GPIO, UART, SPI, I2C, ADC, DAC, and PWM peripheral support.
Full debugging capabilities including breakpoints, source stepping, and variable inspection.
Support for Visual Studio 2017 and the new Roslyn compiler with an easy install.
Practically unlimited max allocation size eliminating the need for LargeBuffer and special handling of large bitmaps.
Improved deployment and startup times.[/ul]
While this preview includes the core features of TinyCLR OS, there is still a lot to do. Here are a few of the things in progress or under consideration:
[ul]USB host and client
File system
Graphics
Secure networking
Device updates
Signal capture and generator
Generics
Controller Area Network
Runtime Loadable Procedures[/ul]
Starting with this release, we will be providing more frequent updates – we want you to be part of making TinyCLR OS the best operating system for embedded devices!

TinyCLR OS 0.2.0 Download: http://ghielectronics.com/downloads/TinyCLR/TinyCLR.0.2.0.zip
TinyCLR Release Notes: https://www.ghielectronics.com/docs/350/tinyclr-os


#2

Exciting, exciting … :clap: :clap: :clap:


#3

Generics! Oye Yea!


#4

(just want to state that I’m basing my judgement on the first release of TinyCLR…if things have changed…feel free to correct me.)

I still feel like .NETMF is the wrong direction to go. .NET Core would have been a better to implement…or at a minimum…re-work TinyCLR to support .NET Standard (so recompilation of 3rd party packages isn’t required.)


#5

Congratulations. Looks like solid progress.


#6

Completely agree with cyberh0me, but if you mean “build with the dotnet.exe” tools, that’s perhaps not out of the question. If you mean “run the full .net Core stack” then yeah - not feasible.


#7

@ User 7412 - awesome first post. You might want to complete your profile or post under your actual account if you don’t want us to think you’re just a troll.


#8

I have always suspected you being a troll :whistle:

Great progress with TinyCLR, well done!


#9

@ All

Does it make sense to continue considering the type of processor that can feat TinyCLR, knowing that WinIOT compliant procs do not require anymore to spend a lot of money to build a board than it was few years ago…?

Is the ‘OS’ based hardware really a problem ? I dont think anymore…

Have a look at the Lattepanda; maybe a board coupling WinIOt and a gadgeteer would be a good competitor instead of an arduino that require another language to be used…
An it will answer the lake of IO that some projects need and dont find in more powerfull Arm procs…not even in Intel one…

What do you think about that ?


#10

@ LouisCpro - this is a very long discussion, dont get us started :wink:

The point is that there will always be smaller and smaller devices that is not feasible to run an operating system on. Think needle with pressure sensor for optimal sewing. Run a full OS on that would not make sense. As long as you have the space and power to run a full OS, go do that.


#11

My personal view on this: I LOVE Win10 but if you are putting Win10 in a coffee maker then I want to slap somebody :slight_smile:


#12

Agreed. Coffee beans taste much better. :wink:


#13

You put Java in a coffee maker. Duh.


#14

In all seriousness, though… As we move more towards voice driven UIs, I don’t think its outrageous to start seeing Win10 (w/ Cortana) in devices of every size. However, I think we’ll eventually come full circle and develop specialized chips to handle the voice UI services next to a typical smaller micro but that’s probably a few more years down the road.


#15

@ ianlee74 - If computing resources continue to grow smaller and cheaper and more powerful and less power-consuming, there really wont be any distinction, I think, eventually. Everything will be an “operating system”. More of the code will just be dedicated to interoperating with other smart devices, down to the smallest sensor level. So you have 1% of your code dedicated to the pressure sensor in your sewing needle, and 99% of it dedicated to allowing seamless communications with anything that wants to communicate with it, it wont matter. But that is a ways off, I guess.


#16

I’m sorry, but I really don’t believe that. If you try to make an OS (or frankly, any software) that tries to work with equal utility on everything from $2 chips up through desktops and servers, you’re compromising somewhere. Java and the .net family are the closest we’ve come, and both of them make trade-offs to pull that off - and trade-offs become substantially more expensive the more constrained your HW platform is.


#17

@ njbuch -

You re right. But in this case, dont you think a couple of lines of code is enough on a pic to make a sensor communicate wirelessly or on a serial line with a more important system…and a pic is known to be far less powerful than what is needed to run netmf…and with less power consumption…

What I see is that I dont know where to place netmf on that scale between a pure sensor and a complex IoT solution. … That was not the case when I started developing on netmf but as Win10 Iot Core appeared, it changed my mind !!!


#18

@ cyberh0me -

Should be interesting if you can detail a bit which capabilities are missing in Win Iot and avzilable in netmf ?

Maybe an explanation is there ?


#19

@ mcalsyn -

Of course it is not a valuable solution to put an OS on a $2 chip.

My opinion is that sensors will require less and less resources and code to manage inputs, and actuators as well, but the ‘brain’ of the solution will represent an amount of treatment, services and memory that will justify to go ahead on an os based infrastructure.

Is there a place for someone else between this two points ? I m not sure…


#20

To came back to TinyClr, it could be a good option to have tool to update firmware in next version (without installing netmf).

Another suggestion, can we expect to have fez raptor support it ? It would be very nice !!!