Lan Firmware Updating, the conclusion

Sounds like you found a commercial/open source product niche to fill :wink:

What it comes down to is that neither GHI nor MS is supplying a product to “the end user”, they are supplying products to OEM manufacturers, ie us.

Also keep in mind that some hardware has NO way to upgrade the firmware, like my buggy piece of crap car radio.

1 Like

[quote]What it comes down to is that neither GHI nor MS is supplying a product to “the end user”, they are supplying products to OEM manufacturers, ie us.

Also keep in mind that some hardware has NO way to upgrade the firmware, like my buggy piece of crap car radio.[/quote]

One has to keep this thread in context. This is applying to LAN enables devices only.

As far as the OEM part. It does not matter who they are selling product to. Either way it should have the foundation to be able to be updated easily by a non tech person. They should not have to install a tool set to update a device. It should be as simple as run the app, select the device from the list, choose the file then click update.

My wild guess is that only a tiny fraction of NETMF devices are LAN-enabled. Certainly overall, the percentage of microcontrollers in use that have LAN connections is miniscule. To invest a bunch of time and effort into a LAN-specific update scheme wouldn’t be a good use of time.

In NETMF 4.2, we got the underlying system we needed to implement whatever update scheme we like. It would be quite easy to build a generic LAN update system on top of that, and it could be donated back to the NETMF project.

@ godefroi, " To invest a bunch of time and effort into a LAN-specific update scheme wouldn’t be a good use of time." i disagree.

Take for instance this case, and point. A engineer is developing a new project that needs to be lan enabled. So he starts cruising the new for new modules out that he can incorporate into his device. He comes across 2 company’s that have modules that fit his requirements as far as flash, ram, i/o & speed.
Company A , which we will call .netmf product, and company B which is generic.
Both products are almost identical in every way, cost, size, flash, ram, i/o & speed. The one thing that separates them from each other is the way you update the firmware.
Company A method is cumbersome and not at all user friendly. Though they have a basic implementation of it which the designer can write an app to do this it would require more time & cost for the project. Company B has a no install app that bangs the network lists all the devices out there and allows the user to click the firmware file and update the device.

Now which company do you think they will go with ? the one that will require them to spend more time and money writing the app, verifying it works on different OS’s, and using the customer as the guinea pig to work the kinks out. Or simply go with Company B who already has it.

Another trouble here is that with the internet age, this could have happened so many times that a engineer came to a .netmf site, liked what it had to offer, but the firmware updates. so he just leaves the site and get product from Company B and you never knew that happened. How much sales are lost due to this, you will never know.

I’m not an EE, and I do this for a hobby, but my general impression is that if you went with non-NETMF, then you’d be doing it completely from the ground up. There’s nothing in CMSIS that relates to IFU, I don’t think. NETMF is already miles ahead of “company B”.

It seems many people here do it for a hobby. I am trying to offer suggestions where .nermf is lacking so it will be noticed more by EE’s.

Actually, its the other way around. Don’t get me wrong i believe in a few years .netmf will be a great thing, IF they do it right. As it sits right now there are quite a few better alternatives.

do you mind sharing the list of the other alternatives that you think are better then NETMF…

i would like to investigate them to see how they fair with netmf…

thanks.

@ Jay Jay, and to others.

My intent for starting this thread was to inform .netmf people that there is a serious feature lacking. Which many large companies consider a huge thing. I want to see .netmf grow. My comments are not to put down but to help. Though i think it has not been taken that way.

This thread has contorted into something else from its intended purpose. I have attempted many times to let the community know the shortcomings but they have not taken heed to my suggestions. There are many simple solutions they could do to grow leaps, and bounds but they neglect the simple things to do. They seem to be more focused on making dozens and dozens of modules instead. When they should have been more focused on laying a solid foundation first. I think the G120 will be a great start if they do it right. Lets hope so, many eyes are watching and waiting.

For Jay Jay, the best embedded module is by far is http://www.netburner.com/
Their entire solution from software to hardware is unbeatable. Been using them for over 10 years now and its solid as a rock.
Then for the linux guys there is the beaglebone.: BeagleBoard.org - bone

Hi jdal,
thanks for the list.
Don’t get me wrong i do share your concerns, as of now my Development has been on hold for over a Month dues to the lack of solid working foundation…which can be frustrating at times… so yes i do agree with you there…which is why we need a solid built solution first instead of just having x amounts of features, and i believe that is the main focus now.

What i hope is this:
I hope the NETMF team had learned how quickly NETMF is involving (especially now) and have been switching their development strategies to adapt an Agile system (Agile software development - Wikipedia )
The same applies to GUS & CO… it is taking them a long time to switch from 4.1 to 4.2 so let’s hope future upgrades (4.3) will happen in weeks and not Months like it is the case now. (which i believe coincides with the introduction of gadgeteer and 4.2) so it’s a lot to do i guess…

Cheers,

We designed and produced about 100 products in very short months. Which is an amazing job from the GHI team if you ask me. But you are right, this is what caused the 4.2 delay. Not to forget the open source offers which we finished with no much help.

The good news, no more waiting. It is all coming together nicely.

My comments are all positive criticisms. Yes it is amazing how much hardware was pumped out in such a short time. But it is also noticeable that GHI spread itself to thin in doing so. Personally i am not a fan of open source. I prefer that when i buy a module all things like SPI, I2C, networking and so on have been thoroughly tested and working. People in my position have no use for a board that is open source where we have to wait for a driver due to the open source nature.

I know the engineer has the option to write the drives themselves, but we have more important things to do such as getting the product to market. The more time we spend on coding drivers and testing them the more the project cost, and in this competitive world where time and money are king there is not much use for anything open source where it needs more effort to get to a final stage.

This is why my eyes are focused on the G120. Correct me if i am wrong but that will not be open source. so from what i understand is that when its released everything will be already working as GHI has written all the peripheral drivers and tested them already. Its a module, so one can plop it into their design and go.

The one simple thing that GHI can do to help it’s presence is to start making detailed videos. Videos from a newbies perspective with no assumptions other than the fact the person watching knows how to use a PC and has some experience in coding. Taking them from what tools to download, and installing. Then opening say a G120 box, making sure it has the latest firmware and updating it. then coding a simple app. Then a other videos by connecting modules to kits and writing aps for them so people can see them work. Video is magnitudes more powerful that all the .pdf’s you can pump out. These videos should be on the main page with text saying something to the effect of “Come and see just how easy .netmf is” they can simply click on the video series, watch them, and i am sure will be hooked.

The bottom line is this. To know if your doing things correctly from an advertising / learning stand point. Is to find someone who knows how to code , One who is not not familiar with VS and C#. This person also should know nothing about GHI and its products. Then sit them down in from of a computer and give them the GHI web site address, and a EMX Development System. Then say to them your task is to get the board firmware updated, and write something simple like “Hello World” on the LCD.

Doing this will show you how well you have done your job. The problem many companies have is that they cannot think outside the box or think like a one who knows little to nothing about them when they make the documents.

Again, this is all for helping GHI improve. When some engineer bumps into your site you have only like 30-60 seconds to keep them there and make them want to know more. A simple video can do this. 14 .pdf’s will not :wink:

1 Like

Thanks for valuable feedback and by the way you see how TinyCLR website is all new but not GHI’s main website? That is exactly what we are working on now.

You can safely assume that the past year was dedicated to open source and to hobbyists/education, not much for commercial users. I think we agree that GHI nailed it down pretty well, from Cerberus and cerbuino to hydra and tons of modules. We are very happy with what we have.

Ok but what about commercial users who do not care for open source? Maybe not even care for gadgeteer?! It all starts with G120 and an all new GHI website. G120 is only the tip of the iceberg coming specifically for commercial users, modules/mainboards/libraries you name it! You being a commercial user, I think you will be very satisfied with what to come, and if not, just let us know what we can do to improve.

2 Likes

sounds all good.

Looking at this from a slightly different point of view…

I am a professional .Net programmer, but I’m also a hobbyist using GHI boards.
I love the concept of open source, and I think it has delivered some amazing software (especially from the guys on this forum).

I chose to use GHI boards because:

  1. They offer a wide range of controller and accessory boards.
  2. Their premium firmware is comprehensive and virtually bug free.
  3. I don’t have to rely on sometimes buggy open source code that is most times limited in features and functionality (as in the Arduino world).

As a hobbyist I want to use GHI boards to do things.
I want to write application level code and maybe a few high level drivers for devices that I connect to the boards.
I don’t want to write, for example, a low level serial driver just to connect a serial device to my controller.

I am also looking forward to the next generation of premium boards from GHI.

1 Like