I was at UCSB when the ARPANET was up in mid 70's running on an IMP processor. I used it to run the MIT Math lab remotely and other things. The Culler-Fried (we called it KFC, but his name sounds like Freed ) on-line system used a card-oriented interface (COL, I think, IBM 80 columns) and a math interface (MOLSF). Had a double keyboard with math functions on top, the output was on a Tektronix Storage Oscilloscope. "Clear Screen" was the erase button! Ran all my Fortran simulations on it connected to a 360/75 with real iron-core memory.

I strongly agree. I developed printers for 3D Systems for past 6 years - they gave up on consumer printers years ago, it just didn't make business sense. It wasn't easy to make the models, FDM couldn't be made reliable for us, so we just stuck with the industrial side (SLS, SLA, Multi-jet, Metal, etc.). Prototyping is the largest market area. I still feel the 3D model creation side is the worse part of it, the software just isn't there, steep learning curve. However, if you like a challenge...good luck!!!