With great interest have I followed the development of the fairly new site mbed.org. It looks fresh, attractive, tidied up.
ARM and its currently 26 listed partners seem very serious about wanting to change how embedded software is developed and devices are connected in the future.
I fully agree with their statement that it is time to bring embedded development up-to-date. They aren’t the only ones who noticed that while say around the year 2000 an embedded software IDE and a commercial software IDE were not that different in functionality, automation and comfort (okay, the commercial ones were slightly ahead), there is a huge gap today. Free IDEs like Eclipse are more productive, flexible and innovative than costly embedded software IDEs from traditional tool vendors. The latter usually run on a single OS only and hardly allow collaboration on an open source basis. We shall be glad that most of them allow for a connection to a version control system. Not to mention that each vendor provides its proprietary library stack. If not more than one. Incompatible of course. (Draging on longer here doesn’t help, thus I’ll stop it.)
Back to the great new idea by ARM and its partners.
An additional high-quality well documented free of cost RTOS is always welcome, and web servers that talk all these machine to machine protocols and allow for the usual front end services may be useful too. (Or not really, I’m not sure if they are necessary, or if it’s just a try of ARM for generating new business.)
Anyway, I opened an account and followed one of their example projects for the STM32F401 Nucleo board. Worked like a charm. The one question I had got promptly answered by the forum community. (The whole project was just about blinking an LED, but hey: it’s a start!)
Do the days of the embedded developers get brighter then? Hopefully, yes! No more days in driver hell, not alone anymore with tricky timing issues on SoC level, comfortable free open source or web based IDEs and easy development, testing and deployment processes. Carried by an open, collaborative community of skilled programmers.
…. …? ..?? .??? ???? ???! ??!! ?!!! !!!!
My dream just ended. Looking at the website again and reflecting its information, it’s gonna take years in my opinion before mbed.org can provide significant aid to industrial production code. They’ll have to answer a series of substantial questions before they’ll get entry into the companies’s development departments, not least what their business case is.
Meanwhile, it’s a neat entry point to the programming world of ARM based controllers, and it may be suited for rapid prototyping. If a project crosses a certain commercial interest level however, my conclusion is: throw the code away and start from scratch.
But hey: it’s a start!
The most uncomfortable thoughts are usually the most valuable ones.