I think there is some confusion here.
We are talking about desktop PNP's being sold on the open market (ie, retail) , not Kickstarter campaigns or any other prototype buying / investing site. PNP's are not rocket science. Yes they take some technical understanding to design / build but most of the China machines are mechanically sound (within reason). The problem is software / support. We aren't asking for perfection here either. We all understand there will be / can be small issues. That is normal. How these issues are handled is the issue at hand.
When these machine builders sell machines with obvious major flaws, with no intent on fixing / providing an upgrade path to unsure customer satisfaction, results like this thread appear. Mechanical aside, spec'ing or designing a controller that has sufficient I/O to support the machine is relatively simple. As long as that part of the design is sound, then software updates should be able to fix most, if not all other issues. In the end, software is key. Most of the machines run Windows based GUI's, which makes the ability to upgrade easy. Why support and the ambition to further develop software and fix bugs is outside the scope of certain machine builders is beyond my comprehension.
For a living, I develop software. As a result, I need to develop hardware. Hardware design, albeit tedious, I find rather cathartic. A lot of the hard work is done for you in today's world of highly advanced IC's and MCU's. Most of my boards designs take less than a few weeks to go from concept to prototype. The software however is a living entity. It is almost always in development. I try ensure software is bug free (that I know of) upon initial sale of the first unit. Software development continues to add features, enhance performance, and if necessary address any bugs found internally or by our customers. Yes it costs more to continually evolve your software, but if you want to survive in a high tech world, you evolve... If a customer purchase fails to work as advertised, I either ensure that an software / firmware update fixes the problem, if not, and I have a part that resolves the issue, I send it to the customer with full credit for the old item, or I refund the customer. I can't afford to have unhappy customers. It is far too costly to obtain customers. It is much easier / less costly to keep them happy. I also find that the cost of ongoing development actually costs less in the long run, than trying to support work-arounds to bugs.
I still don't understand why certain companies manuals / documentation is so bad. Really hire someone that knows nothing about your products, and write down the answers to all the questions they have to make it work. For the most part. Your manual is now complete.
Sure, maybe a few of us want to buy a piece of hardware and get the protocol to write our own PNP software, but in reality, I think most of us just want to buy a machine, at a certain price point, with certain features, that works as advertised. Essentially, take delivery, setup, and build boards.
Maybe I am wrong. But that is what I wanted / expected....
Desktop PNP market is still for early adopters. Perfect example/approach is by Axel/servokit who is happy with a mostly working machine out of the box and spends endless time tinkering/upgrading. I have been watching smallsmt developments since the beginning and would not be worried about purchasing from them. Platform evolution/refinement has been very impressive. But my expectations for any machine are relatively low because that is realistic for this market at this time.