OK Janie Love, this is very painful for me to type, but I have to agree with your article: Custom Hardware Design is Dead. On the other side of the fence, I am going to say that custom hardware design is not dead yet, but it is in critical condition. As long as people are still using it, it is alive and kicking. It only takes an article like this and comments like the one from Eric Starkloff, senior vice president of marketing at National Instruments to motivate a team to get busy. We will see.
The reconfigurable hardware from NI will definitely a big hit where ever we have a less effect of the hardware price on the total solution offered. It is going to be a very big market for the NI in next few years without any competition.
Platforms are excellent enablers of complex technology solutions, but one of the key elements necessary for innovation and new design is the ability to quickly turn an idea into reality at the hardware level. In the past this was possible by hand wiring SSI integrated circuits in low pin count packages or using a solderless breadboard with jumpers. This approach is still possible but it has limitations. I have recently been using a new desktop programmable mixed signal ASIC chip from Silego that solves the low cost, small size and production manufacturability issues. I now have true desktop prototyping and real time emulation for less than $60 and I LOVE it!
Yes, it is undeniable the popularity gained by platforms, but at the pace technology changes, how would new platforms be developed without custom design? I have a proposition for NI guys: remove all custom design from current and future platforms to see how much growth they will experience.
Dr. Truchard's talks usually have hints of that to come from NI. But if some kind of common hardware platform emerges, it would likely ne Aruino, Raspberry Pi, or something like that. For saftware development, the "common" platform is iOS and Android. But, one size will never fit all.
At the first NI Week (1995) NI gave out T shirts. On the back was "Top 10 Reasons why I attended NI Week 95." One of the reasons was "Rally en masse for Undo in LabVIEW."
A few years Later, NI Week was big enough to move to the convention center. NI gave the usual snappy keynote. Near the end, they made a coding mistake and quietly used undo.
When I tried to write about it, my print editors shot me down saying that it would do nothing for people new to LabVIEW and I should focus on features that would attract new people to it. Times have changed. At that time. we would print "LabView," as opposed to the proper spelling, LabVIEW. the reasoning was that using all the caps drew too much attention to the name.
VIEW is actually an acronym. Who can tell me what it stood for?
I have always been amused at people who say things without thinking first. Custom hardware will ALWAYS be needed, as the requirements change too much to have a standard HW design meet all needs. Even if you standardize on a paticular SOM module, you will still need an interface PCB to connect it to what ever physical devices your application is controlling or monitoring.