Many thanks for all the kind words, both on list and via private email. I'll try a catch-all response to clarify a few things and answer the questions.
The parts in the photos before the prototype shot of the running board supports are all 3D printed masters with Archer rivets added. The parts in the last two photos (the one showing all five parts and the one showing them dry-fitted) are first generation resin castings made from those masters. Those first generation castings will become production masters after I add a few more details and make sure everything fits properly.
From my standpoint as a resin caster the original 3D printed part only needs to get me through the Archer rivet application stage. At that point I'll make a mold and work with resin castings from then on. There are two reasons for this. First, the 3D printing processes that are capable of rendering the level of detail shown in my photos use materials that are much too brittle or unstable for end-product use, and the processes that do use suitable materials can't render the required level of detail. Second, Archer rivets are wonderful, but they *will* chip off with careless handling and, in my view, aren't trustworthy on a master from which 15 or 20 molds might be made. So my designs allow for at least two generations of mold and casting shrinkage, and I use first generation resin castings rather than printed originals as production masters.
Yes, the idea is to market two resin kits, one of which will make a Tk-G or Tk-H, and the other to make either a Tk-I or Tk-J. Each kit will have a tank with appropriate details and two domes. I intend to do the later versions of the Tk-G and -H, with center tank anchors, rather than as-delivered with head blocks and diagonal tank hold-down straps. The diagonal strap anchor castings on the top of the tank will be there, of course. A few things have to be worked out - the running board supports, which in this design are cosmetic rather than structural, and the running board steps might best be made of brass, but whether from formed strip or photo-etching remains to be determined.
As Kent Hurley indicated, these parts were not made on $1600 machines. Or on $16,000 machines. As many of you know, in my post-retirement career I consult part time for a local rapid prototyping company and have access (but not free access) to some pretty amazing technology. We can make parts using stereolithography (SLA), selective laser sintering (SLS) and PolyJet. We don't yet have any MultiJet machines (like PolyJet but with higher resolution), so I use Shapeways for that technology. At this stage of their development I wouldn't consider buying any consumer-grade 3D printer. They do have a lot of gee whiz appeal and are great for making decent 1:1 objects and shapes, but in this hobby we make miniatures, not full-size objects. (In the trade, the consumer-grade 3D printers are called "computer controlled hot glue guns".) Pay attention to the jewelers who use 3D printing. Their needs are similar to ours, so follow their lead.
Now if you will indulge a bit of philosophy…. I am a competent modeler, have become passably competent in 3D CAD, and have expertise in resin casting and mold making. When it comes to casting I can look at a part, visualize how the resin will flow, where the parting line needs to be, where the gate and vents should be located, and how the part should be oriented in the mold. I've also been intimately involved in the design of a very complex product line, the Branchline heavyweight Pullman kits. So I have a pretty good view of the whole process, from concept through manufacturing and final assembly. My gripe with resin kits in general is that the patterns are created but not designed, and the kits themselves, the aggregate collections of parts needed to create models, are seldom engineered with any consideration for ease of assembly. That's understandable because, traditionally, patterns for resin casting have been created by expert scratch builders. And scratchbuilding involves a lot of "make it up as I go along". Also, even with one-piece bodies, many parts are still made by flat casting and you can't include mounting bosses or alignment keys on the backs of flat cast parts.
Tank cars are considered very difficult resin kits to assemble. I've tried to break a complex model into parts that are straightforward to cast, and include registration and alignment features. All those notches, grooves, posts and apertures in the parts have a purpose. Drill and tap the bottom of the dome mounting boss and the subassembly of upper & lower tank shells, running board and dome can be held together in perfect alignment with a single 1-72 machine screw. The tank supports that mount on top of the underframe ("fiddly bits" awaiting application of Archer rivets) are keyed to it, and the underframe is keyed to the lower tank shell. Holes will be spotted for grab irons and handrail posts, and a forming jig provided for making the offset bends in the end handrails. You'll still have to remove flash from the parts, but once that's done the assembly should go quickly.
Might be a Prototype Rails clinic in all this - Designing for 3D Printing, including an overview of the 3D printing processes appropriate to our hobby....