@AZskibum, no, i don't think that a chip company would rely "solely" on firmware, drivers, etc. developed by the open source community...but it would help, I imagine, the problem of having to handhold so many potential customers by fielding all the questions.
As somone said, "Friends help friends for free" is what chip companies are dreaming as the purpose of their so-called "community," is it not?
Well I'm glad that someone else feels this way. I constantly find myself trying to explain where the maker stuff fits. I do a lot of contract design for smaller companies and it's not uncommon for the CEO to proclaim that he is going to replace everything with Arduino.
Never mind that 90% of the product is custom power stuff. Typical wishful thinking I suppose.
I also run into a lot of younger less experienced folks that think the only one who makes microcontrollers is Microchip. They invented it right!
I try to explain that they might be better off using a $99 board from Freescale rather than trying to splice all sorts of Arduino stuff together. It falls on def ears. Somehow free source is not as good as open source. C is scary and sketches are not.
On the other hand, it could take days to find that $99 demo board on the Freescale website and even at that it might not be obvious that it is actually what you want. It is also not all that easy to figure out that the free versions of their development tools are hugely productive high quality tools. What a deal!!
I am sure that the same is true for many other vendors. The issue is that for me this particular one is easier than Arduino for virtually everything I do. However it has taken some time to tease that out.
If a manufacturer can gain some visibility outside of the small group that already knows them, so much the better if it's not a huge investment.
Good article. I think there is another angle here.
Back in the 70s, the standard success story was 'management didn't want to take risks on eating their current business, so we struck out on our own with an idea". Not a problem, since HP, DEC, and IBM ruled and the pesky startup types were just a pain in the management's butt anyway. No Prez will loose his job.
In the past few years, the opposite situation has come up. When Intel turned down Apple's request for a processor for iPhone/iPad, the storm cost people their jobs and the company a good chunk of their capitalization. Every BoD now hammers Exec staff on how hip they are to new types of customers. No chip company Prez will survive without a 'maker' story to tell the BOD. "Yes, we are building tools for the guys in dirty Hoodies who hack IoT stuff!" No one on the BoD is asking about ROI questions about the unwashed semi-masses yet.
It is easy to put a small team in place, give some love to the open source community, and issue a bunch of Press Releases. Not so it you have to go to the BoD and explain why you missed a market.
Here's another factor. In many large companies, there is Business Area One (IBM terminology), which is the largest money-making part of the company. For IBM, this used to be mainframes, for Intel it's the x86 line. Typically, the management of Business Area One calls the shots. After all, if BA1 fails, the whole company goes under.
In some companies, if you're not in BA1 you're a second-class citizen. If you're a chip-maker, the silicon guys are BA1. So software and making evaluation and development boards could easily end up having second-class status and the company will have trouble hiring and retaining good people for those jobs. In that case, outside software and eval board developers could end up producing far better products than the company itself. If a chip has exciting capabilities, people in the developer community will get excited and create wonderful software and boards, provided that the chip-maker publishes adequate documentation. "Of course, the whole point [...] is lost, if you keep it a secret! Why didn't you tell the world, eh?" [Dr. Strangelove (1964)]
ASskibum wrote about: ...software that the company itself cannot support...
As long as the community-developed software uses a standard open-source license, the company can at any time fork the software and support the fork themselves. If it's a copyleft license like GPL, the company will need to release the improved source code with binaries. If it's a permissive license, they don't ever have to release source code again. (At least that's my understanding, IANAL).
The community has done the hard work of exploring what the software should do and figuring out what tricks are needed to get around the quirks of the chip, so those high-risk, high-cost (in terms of programmer time) factors are taken care of.
A great way to bring the software in house is to hire the key community developers, which gets the company immediate expertise and demonstrates that yes, programmers can make real money writing free software. Plus, if the key developers wrote all the code they can relicense the software to the company on a proprietary basis.
The maker community's collective efforts to develop firmware, drivers, etc are great, but do you honestly think a chip company would rely solely on that -- software that the company itself cannot support -- as the only means of enabling its reference design?
One interesting place to look at the effect of maker communities on real products is the mbed community: it's relatively old, uses commercial friendly licenses and is relatively easy to transfer to manufacturing. Not sure how to get the data. Maybe asking mbed or arm, the company behind the project, is a good way.
A Book For All Reasons Bernard Cole1 Comment Robert Oshana's recent book "Software Engineering for Embedded Systems (Newnes/Elsevier)," written and edited with Mark Kraeling, is a 'book for all reasons.' At almost 1,200 pages, it ...