Can't we just use a Pi?

I love the Raspberry Pi! We use them all the time for proof of concept or solving problems quickly with minimum hassle and stupidly low cost. In fact, I always have one in my work bag just in case I need to knock something together on-site. More recently a customer gave me an idea for production test. Rather than have one test computer that holds a number of setups for different customers products, have a 'test' Pi for each project. Each bed of nails comes with a Pi, just plug in a monitor, keyboard and mouse and away you go.

In the last few years more and more customers have said the immortal words "Can't we just use a Pi in our new flash product?" (other development boards are available). That used to be an easy question to answer - "NO" for the following reasons.

• Most boards didn't offer an extended temperature range.
• A lack of commitment to availability.
• A lack of commitment to design compatibility
• Potential sourcing issues when huge orders are placed. The Pi foundation actively discouraged large orders (this has now been completely relaxed)

So up until now our guidance has always been to steer people towards the use of SOMs (system on modules) which are generally more suitable to commercial products. In fact the Pi Compute 3 module is a very competent piece of kit and is modelled closely on the standard Pi 3. Even if the project starts on a Pi 3 it can ported over with ease to a module at a later date. We have used SOMs on many projects (both Linux and CE) to great success, but it's not always the cheapest option (in terms of component price at least)

The Beagle Bone Black Industrial can now operate in temperatures from -40C to +85C, which is fantastic, High profile development modules such as the Pi and BB are standing the test of time. As the longevity and robustness of these development modules continue to improve, I believe we will see more temptation to include them in medium cost, low volume designs.

At the moment, they still make me a bit nervous, for use in a customers project at least, not because of the problems today, but the issues down the line. The counter to this argument is fair: there are enough people doing the same thing, that a sensible solution will be available.

Does any one have any first hand experience or views - good or bad?