So I’m playing around with my Medusa Mini for 2 days now and I kind of ‘hate’ it already.
Compared to NETMF devices it’s like hell!
1st: No debugger:
I tried Visual Micro, for VS, but the debugger does not stop on breakpoints >:)
2nd: Suddenly stops working.
I have a running program, I do not really change anything (just cleanup), recompile, deploy, seems to stop in the middle of program.
Power off, different board, restart VS … nothing helps.
So I add a couple of Serial.println (RS485 is disabled to do so).
With that everything works again.
Removing the println … still working. ???
This really reminds to the programming lessons in university:
Linux, GCC, some (very simple linear) program does not work … add some println -> works … . remove println, does not work :wall:
By this, I do not understand how any sane person wants to use Arduino.
I do also not wonder anymore why there are no powerful Arduino boards: You never can write a program that is complex enough to use much more resources.
And I can imagine using the Arduino IDE is even worse.
This does not speak against the MedusaMini: For my application the board is perfect. Just the development for it is really really hard
Arduino is great but has its own place. Thousands of users religiously love it but you have to know when to use or not to use an 8bit micro with no debugging. It is still a $10 board so it is a good option for many simple things. If cost is not an issue and you do not have real time needs (native coding) then NETMF is probably always better for you, in my own opinion!
@ Gus - Certainly it’s much more complex. Getting Eclipse (for example), your JTAG server, GDB, etcetera all connected correctly is a chore, and there’s little help for someone unfamiliar with the tools.
But it IS possible, and it IS powerful, and it IS native code. There are still things that the lowly 8-bit AVR at 16mHz can do faster than NETMF at 400mHz.
@ Gus - I have got nothing against any of these technologies, but this is what you should do
[ol]Create a small JTag emulator, here is a open hardware one http://www.coocox.org/CoLinkGuide/CoLinkDIY.htm
Create a module with the LPC1113/4 chip on it with Gadgeteer sockets
Port your Medusa library to the LPC1113/4 chip
Use CoIDE,EmBlocks or one of the other free open source IDEs to progarm [/ol]
This will give a native programming environment for Gadgeteer with a great ARM chip with full debugging and I am sure you could do it relatively cheaply.
Sorry, I cannot help but promote the LPC1113, the DL40 that you have is my favorite module that I think is way under used. I realize that you cannot use all the cool Arduino libraries that are available, but realistically the people using Medusa seem to all be using the standard library modules anyway.
Contentions point I am sure, but a DL40 with gadgeteer sockets would be perfect!
Backward compatibility becomes a hindrance to innovation. Myself, i would not focus my development efforts on what some other company designed several years ago - I put my money on doing the next thing, not the last or even the current thing. I would rather not see the focus of Netmf shift toward Arduino compatibility. I would rather see new gadgeteer modules with more functionality than try to use an existing shield build for a different technology.
I’m sure Arduino has it’s place. It’s a low cost hobby electronics kit for relatively low requirements.
To my current experience I would not use it in a commercial product, because it’s hard to find out what’s wrong, if there is something wrong.
If you think on “quick to market” (for professionals) or “low frustration level” (for hobbyists) I do not think that Arduino can compete with NETMF. Since NETMF boards are also more power full the only thing that speaks for Arduino is price and it’s “market share”.
I have to agree with Blue Hair Bob. Backward compatibility, usually in the name of the bottom line, holds back innovation. Arduino will become antiquated, Gadgeteer will eventually become obsolete, and even silicon and copper will fall to the wayside as innovations replace them (optical computing, anyone?).
I agree in the way that backward compability should not stop innovation. It nearly killed the x86 platform (about 10/15 years ago!?) because of too many bottle necks.
Only new innovations starting with pentiums and PCI solved the issue (kind off).
But I don’t see the NETMF platform at it’s end.
Arduino compability: Not really needed, but may drag some Arduino guys to NETMF.
But I do think the NETMF core might need some fresh air within the next 5 years.
So what does NETMF stand for?
Its a .NET based, minimal operating system for µC’s.
So a new core needs just to be that.
There are several possibilities to improve the core, some are already discussed in other threads, like compiling IL to native code instead of interpreting it, a pure managed OS like COSMOS might be one some day, …
If COSMOS could be made to run on small hardware, say Pi size/cost, I think that would be an enormous win for the .NET community. I’d probably back that kickstarter project.
For NETMF, however, I think the number one huge missing feature is native code compilation, or, at the very least, a significantly better native code interop story. I’d back that kickstarter project as well.