Your feedback to Microsoft for NETMF and gadgeteer

Archive
21

Pay Oberon not MS to extend the STM32 port, they are the experts …

4 Likes
22

I would echo the comments about documentation and examples. Being a newbie to NETMF and Gadgeteer, it takes a lot of reading to figure out what is the LATEST or best way to do things as the Framework and hardware evolves.

The Gadgeteer Home on Codeplex points to this blog (http://www.netmf.com/gadgeteer/) as the place to go for information but it seems to be rarely updated and rarely used.

To get people excited, you want to be able to get things up and running quickly. As lame as this may sound, you want to be able to see the flashing LED (aka Hello world) work without having to figure out what I need to download to make this work, why I cannot compile or why my firmware wont update or my USB cannot recognize the device etc.

While I understand there is a learning curve to be had for a more complex solution, to be honest, it was a lot easier to get up and running with Arduino than NETMF/Gadgeteer. There are just so many more examples and information.

In addition, make it easier to get the devices. Almost everyone who is into microdevices sells Arduino but only a few sell Gadgeteer. I had inquired about reselling twice now, but have gotten no response from GHI. But I did get quick responses from Arduino and Mikroelectronika. Speaking of which, THERE is a great example of a website to emulate in terms of examples, forums etc.

Despite all this I decided to go with Gadgeteer for our company because of my familiarity with Microsoft products, programming and because it just looks that much more professional. But I have to admit I was very tempted by Mikroelectronika…

1 Like
23

See this post that was added not less than 45 mins ago is exactly what I am talking about.

https://www.ghielectronics.com/community/forum/topic?id=15469

Specifically about the installation issues and more generally about the guides etc.

24

Non-Volatile Settings: I want the Settings class from regular .NET to work with Visual Studio and persist to the flash on the MCU.

Wifi: Robust, inexpensive, reliable, and feature-rich Wifi including HTTPS and WPS.

USB Support, I want to be able to read a FAT32 USB stick and I want to be able to reflash a device by inserting a FAT32 USB stick containing a firmware file.

Microsoft should pick 2 processors (small/cheap and bigger/faster) to focus on and implement reference designs with reference source, firmware, drivers, and web server demo app that allows you to turn on/off outputs and read inputs from a web page. I know the porting kit has a few already, but I want one for something like the Cerberus. The Cerb Family has a fantastic variety of form factors and yet a single firmware for all.

Porting kit should be GCC first, above all other compilers.

I would like to see a completely managed C# version of MFDeploy.

And an open source C# application that allows you to build custom firmware with different options. I know this was tried before, but it deserves another try.

1 Like
25

Indeed. NETMF, while open source, was clearly designed to enable companies like GHI to build hardware solutions. It was definitely NOT designed to allow hobbyists to build firmwares.

Assuming there was a native compiler (much more likely if there were only one platform targeted), there would be no need for a “bigger/faster” solution, especially considering that the STM32 hardware now has support for DRAM and LCD.

26

If they do AOT, JIT or .NET Native, the compiler framework should be modular.

(assumption is that processor specific code would need to be used)

Not sure if MS will invest the time to do the work for each supported processor. So when STM releases a new ST32F4?? - the community could do the leg work.

Some type of VS native compilation would allow less capable chips to be used, which in turn would lead to less overall expense. This would put it

**

Also I agree with others a greater presence in the community, through marketing/adverts, support, continually updated official sites and more visible demo projects. The piano at Build was meh, for me.

Their official site should be tutorial and demo rich. Right now they are too dependent on the vendors. That site now only has 5 demos and a simple list of links for resources.

27

ADD DNS Server
ADD DHCP Server
ADD WPS (WiFi Protected Setup) support

In general:
Any IoT device needs easy setup capabilities for "dummy’ end users

28

Along those lines, I’d like to suggest to Microsoft also that they extend the MVP status to NETMF/Gadgeteer. That would be a great way to create more community champions and costs them very little to implement. Geek cred is cheap and a very effective motivator.

1 Like
29

@ andre.m - A DNS server sounds heavy. What i mean is a system to access a device by name instead of IP address.

30

Let me just add to the choir: powerful networking and low power. Period.

Fascinating that this is still an issue.

31

And “low power” not just means sleep modes, but also compilation instead of interpretation. Executing just one instead of, say, 30 machine instructions not only means a speedup of 30, but also a reduction of the power consumption to about 1/30 for the same program.

4 Likes
32

I’m sure you don’t mean servers… they don’t help you set up anything, and aren’t really useful unless you’re building a router or firewall…

The framework already has DHCP and DNS clients, though they’re not always perfect. WPS might be up to the Wi-Fi solution to implement in most cases.

33

Message from a newbie:
Someone has already mentioned it, but I will repeat it:
Please keep the information in an actual state! I spent a lot of time to read wrong or old information.
Need a special example?
The serCam only stores a compressed jpg file (how I found it out? space on the sd card a a jpg sniffer).
But I need a real bmp-file, because I want to analyze the picture. After 2 weeks I give it up. I have really no idea to whom I can tell the problem.
I think an information technology lives from information…
I use a cerbot with a SDCard, a serCam an a button just for fun, but there is no fun yet.
I can’t recommend this for teaching.

34

Just echoing important issues for me (in order of my preference)

  1. .NetMF website needs more activity, articles, blogs, examples of best practices and code snippets (honestly looks really dead there and is barrier for anyone looking to use the technology)
  2. faster and better issue tracking and resolution
  3. reliable and easy implemented networking
  4. update .NetMF to support new features of .Net eg. Generics(I believe it is coming)/Linq/Parallelism, make switching from .Net to .NetMF and back easier
  5. AOT compiler
  6. support multi cores

I believe .NetMF/Gadgeteer platform is a great platform and has a great future if some of the issues are resolved.

35

Don’t bet on it. In fact, bet against it.

I assume you mean by this something like the LPC4300? An asymmetrical multi-core? Do you have a specific use case in mind?

Oh, by the way, while we’re on the topic, I’d gladly give up threading if it meant gaining an AOT compiler.

36

@ godefroi -

YOU WILL HAVE TO TEAR THREADING FROM MY DEAD HANDS! :naughty: :naughty:

2 Likes
37

Let me add some things :

  1. Be able to use more than one network interface at a time

  2. More improvments on DPWS, and secured transport protocols as on the full framework (IoT)

  3. Interoperability between class serialization on netmf and the full

  4. Ipv6 support,

  5. SNMP protocol inside (agent and client)

  6. Timeout on socket TCP

  7. More powerfull graphics (wpf/xaml)

Quite all…

38

Agreed I don’t wan to be writing state engines.

39

Hi Ingeborg. People on this forum are normally very willing to give advise or help. But this is not the right thread to ask this question as it has nothing to do with the the topic under discussion. I suggest you repost your question as a new topic under gadgeteering and try to be as specific as possible with your questions.

40

Then what’s the point of the whole NETMF? Take a bare metal ARM and you’ll have what you want…