mbed is not new
Been around for at least a year or two.
The cloudy bit might be new though.
ARM will today announce a new operating system called mbed OS: it seeks to smooth over all the differences between various competing system-on-chips so that high-level applications can talk to sensors and other gizmos whether they're using silicon from Atmel, Marvell, ST, Freescale, NXP, and so on. Diagram of mbed OS and …
You got a license for me to sign? Your product dies.
Errrr... And why do you think their product will die just because you disagree with signing licences?
If people producing products didn't agree with licences, that'd mean there were no Freeview or Freesat boxes on the market - Not that that would necessarily be a bad thing, as such, because you could still create a box that'd *receive* the stuff - and it would not be able to decode HD content (Because you need a licence for H.264 which they use for HD) and no Dolby decodes. (And thus some of your channels would be quiet)
Just because you - who wouldn't generate them any money anyway - disagree with signing licences, it does not mean that the people who *do* generate money for them won't.
I've been meaning to get back into embedded systems for a while, I've got silicon from both ST and Atmel kicking around, time to dust off the emulator.
(the IoT could be useful in a business/commercial environment, I'm sure that I can come up with a killer ap or two given some stimulus)
So you say you'll handwave the technicalities of the hardware away, I just have to run some binary blob to do it? Nice try, no thanks. I'm sure I'll be professionally shoehorned into having to use this or something similar sooner ar later at some point, but as long as I have a choice: good riddance, mate.
Except you don't know that.
Looking at the current tree, (A casual look through some of the directory - admittedly not comprehensive search through everything) at the very least the ethernet drivers for Freescale and NXP are available as source under, what looks like, a BSD like licence.
https://github.com/mbedmicro/mbed/tree/master/libraries/net/eth/lwip-eth/arch
There's also a bunch of other drivers under this directory:
https://github.com/mbedmicro/mbed/tree/master/libraries/mbed/targets
Just dipping in, plenty of .c files and .h files, the odd .cpp and .s files. The odd one has the Apache licence.
In the article there is mention that the TLS library is not open-source - part of it, at very least, appears to use axTLS - which is BSD licenced.
The idea here is so that people who are handy with C++, JavaScript, HTML, Swift and other languages for phones, tablets and desktops can prototype and build applications for fiddly hardware ultimately hidden away under the mbed stack.
What? Both of them?
I'm not sure where this idea came from but it will take some work to convince me that the async patterns of the web-twiddlers will work well in the embedded world. But Erlang might see a renaissance!
Most operating systems do need far to much RAM to work on those systems... however offering a C++ API will attract many people who don't know C++ enough to write embedded applications for it. C++ has a few odd features like implicit object copies which make using it on embedded systems rather hard.
Also if you want an embedded operating system, particularly a closed source one, code quality is _much_ more important than features and claimed support, as you'll be spending much of your time debugging the operating system, not your actual application.
If you want an embedded operating system in that area, I can recommend you FreeRTOS/OpenRTOS. It's a small code base of very decent code quality and very easy to work with.
I read only the first five paragraphs before BOOM, they dropped this bombshell -
>However, it will work only with kit fitted with chips based on ARM's Cortex-M designs.
>That means that despite the Cambridge-based firm's stated aim of addressing "fragmentation" in the so->called "internet of things" (IoT) market, the code will not work on devices that use Intel's Quark and Atom >chips or Imagination's Mips processors.
Way to trumpet British innovation BBC! This is a total non-issue given the VAST majority of mobile devices out there use ARM chips. Who the hell uses Quark or Atom outside of the cheap-o Android tablet range, not to mention the entire embedded computer industry.
Thank you, el-reg, for not shitting on British innovation like the BBC seem to do regularly!
Guhhhh! The BBC get on my tits.
The BBC mostly don't annoy me - but this seems particularly bad given that it appears to be BSD/Apache licenced and the MIPs and Intel crowd could fork it and write their own drivers, if they so wanted. (Imagination may be interested, but I can't see Intel looking to something quite that small)
... there has to be a starting point somewhere.
So how does the UK enter in to worldwide market with worldwide players of some established track record and lots of kaboodle in the bank like ms, happle, thegoogl, inttel, ... with the support frameworks and collaborations naetworks in place.
Use ARM?
No!
Set a startup with funding and let Joe Soap at the dole office manage it?
At least one earlier post (I have not read them all yet) identified it has having been around for a while. Wikipedia says September 21 2009. I know I have seen the name on many occasions over the last few years, so it is not correct to call it new.
That minor grumble apart, I think this is potentially good stuff, and those who are sceptical may like to pause for a moment and consider that ARM processors far outnumber any other on the face of this planet. Do not for a single moment imagine that a small, by global standards, Britich company can not conquer the world, because they already have, unseen by most people.
I have no connection with ARM, but I do use an unknown number of their licensed products and at least 5 or 6 that I do know of, Raspberry Pis and phones. I think they are in a couple of gadgets and NAS boxes too, but have not bothered to check.
This market penetration is a bit like free software, including Linux, GNU utils, Busybox, and umpteen other things, mostly produced by individuals or smallish businesses, by global standards. Such software is found almost everywhere, starting with most server rooms, every Android phone (by far the brand leader now, usually also running on ARM), and almost all supercomputers.
But basic commodities like wood, metal, plastics, food, air and water go everywhere too. Just because something will NEVER be a globally leading brand like Coke or McDonalds, does not mean that its uptake will be small or insignificant. A close examination of the facts (i.e. what stuff runs the world and keeps people happy) will show that lots of almost unseen, unknown things actually have massive commercial success behind the scenes.
mbed will stand or fall on its merits, not because it is emerging from the UK. UK bashers and Apple and M$ fanbois take note!