> Like a lot of Arduino products, it must be targeted at people who don't know any better and whose sole exposure to electrical engineering or systems engineering is what they see in make videos.

Industrial automation is exactly this. If you look up the history of the PLC, Dick Morley, founder of Modicon, specifically designed it to hide the computer part. The reasons were twofold: partly for marketing as in ye olde days computers were big, expensive and required costly engineers, and to enable plant electricians and other maintenance personnel the ability to program them using ladder schematics which is why we have ladder logic. So from the beginning the PLC has been targeted at non-technical personnel.

> Automation direct PLCs are cheaper than whatever they're hawking here and almost certainly way more robust (having used a few, they were surprisingly good).

Indeed, though I only like the Click. I have two Clicks at home and used them at work for all sorts of little projects. I haven't bothered with their Productivity line as I dont think it's a very forward thinking platform as its stuck in the 90's with parallel busses on backplanes with proprietary expansion (literally a clone of their Koyo DL series). Then there is their legacy Direct Logic series from Koyo which is solid as hell but long in the tooth. Now they are selling LS Electric which makes me feel they are all over the place PLC wise.

For bigger automation projects I use Beckhoff gear which is very modular and their EtherCAT protocol is very well designed and thought out - deterministic structs over Ethernet. My only gripe is the big shitty Visual studio shell IDE that tends to fall over but every automation vendor has big shitty IDE's. Thankfully they have a FreeBSD runtime alternative to Windows Embedded for the controller side. I can write real-time PLC programs in C/C++ and whatever on the Windows/BSD side. Or stick with IEC 61131-3 languages.

> Anyone who is using PLCs in a remotely serious industrial fashion is almost certainly going to keep buying Rockwell or Siemens anyways.

That is more of an industry trope: "no one ever got fired for buying IBM" but replace IBM with AB for the USA or Siemens in Europe. Basically, Enterprise grade automation with all the associated enterprise goodness: high cost, licensing, price gouging legacy support, and poor technical decisions. I stay away from it.