Comment Re:Hire An Expert (Score 1) 68
That guess is highly accurate. I'm the only software engineer here, and all of the embedded work I've done in the past has been on OS-less targets. This time around, that simply isn't an option due to time constraints and peripheral connectivity issues. The only truly embedded OS I've personally used before is QNX. QNX is great, but way too expensive. In fact, VxWorks and pretty much all of the usual suspects just cost too much, and are overkill for what we need.
Based on cost, board support and driver availability, GUI development options, use in similar applications and V&V documentation, Linux is turning out to be the best choice.
All of my friends who've worked on embedded systems have done so in larger companies -- companies that can afford to staff people to sustain a Linux distribution. Here, "sustaining" can't mean hiring somebody to write and apply kernel patches. We're small now, and will continue to be small (on the software side) for years. "Sustaining" to us means finding a vendor that will provide adequate support for a frozen version of their OS for the next five years.
And that's my worry with Linux. Releases come often, and the vendor I'm looking at most closely has a "Current version plus the previous two" policy. If they release aggressively, that could mean our platform is obsolete in three years. Yes, we'll still have source code, but we'll have to hire somebody to maintain it.
As far as I can tell, there are no experts. There are only people who sell both consulting services and operating systems, and they quickly suggest hiring them to do all sorts of sustaining work. Other people that I've talked to chose an OS because of some specific make-or-break reason, and never revisited the choice.
I guess the ultimate question I have for the Slashdot universe is, "How long do you think an embedded Linux provider is going to stay in business and continue selling Linux-based services?"
I know the answer will probably be, "Forever because Linux is teh sexy", but I'm hoping that some grizzled veteran will tell me a war story about the true cost of software maintenance -- for any of the big embedded OS's.
Based on cost, board support and driver availability, GUI development options, use in similar applications and V&V documentation, Linux is turning out to be the best choice.
All of my friends who've worked on embedded systems have done so in larger companies -- companies that can afford to staff people to sustain a Linux distribution. Here, "sustaining" can't mean hiring somebody to write and apply kernel patches. We're small now, and will continue to be small (on the software side) for years. "Sustaining" to us means finding a vendor that will provide adequate support for a frozen version of their OS for the next five years.
And that's my worry with Linux. Releases come often, and the vendor I'm looking at most closely has a "Current version plus the previous two" policy. If they release aggressively, that could mean our platform is obsolete in three years. Yes, we'll still have source code, but we'll have to hire somebody to maintain it.
As far as I can tell, there are no experts. There are only people who sell both consulting services and operating systems, and they quickly suggest hiring them to do all sorts of sustaining work. Other people that I've talked to chose an OS because of some specific make-or-break reason, and never revisited the choice.
I guess the ultimate question I have for the Slashdot universe is, "How long do you think an embedded Linux provider is going to stay in business and continue selling Linux-based services?"
I know the answer will probably be, "Forever because Linux is teh sexy", but I'm hoping that some grizzled veteran will tell me a war story about the true cost of software maintenance -- for any of the big embedded OS's.