I'm afraid I don't have very much time to frame a reply, and by the time I have another opportunity - next weekend - I shall have forgotten much of what I had to say. Here are a few thoughts off the top of my head, though.

I think your stepping-back exercise was wholly worthwhile. If anything you should maintain that stance even while proceeding with actual applications. There are more questions that need to be asked about exactly what a project is to achieve.

I was reading another thread here about building systems, and I think my response is better posted here than there. I'm an architect and have come to have a dim view of building systems (as opposed to alternative methods that do not constitute a system in that sense), first for their sheer inflexibly and consequent rapid tediousness, and later because of the underlying failure to understand the problem.

In organic agriculture there is a growing understanding that the problem of world hunger is one not of food cost but of the practical power to produce food, and that cheap food is therefore not a solution to it, but may in many cases exascerbate the problem especially where it coincides with further losses in farming skills continuity, land access, etc. Robert Luis Rabello explains it here. This is precisely parallel to the problem of shelter: in precisely the same way shelter is not a problem of construction cost; it is certainly not a problem of construction speed! but very much one of land tenure, access to materials, continuity of skills and technology in a culture, etc. And in precisely the same way the problems of mobility and energy are not ones of cost nor even of resource use ("efficiency") but of the social, economic, political origins of contingent need structures.

This is by no means to imply that an engine project, for instance, has no merit. I for one just don't see it having much to do with a "developing" context. I think it can do the most good in an established Western context by undermining what economists of the Austrian school call "roundaboutness". They consider it a goal; I consider it a scourge. But even here the project will have a political dimension (and may therefore expect political opposition) because established interests derive from "roundaboutness" their economic power, both by their techno-political ability to command large portions of markets but, more importantly, to amplify the scale of those markets out of all proportion to anything like spontaneous growth. It is almost impossible to gauge what spontaneous demand-driven industrial output might be in the absence of economic structures amplified by "roundaboutness". I am by no means unbiased, but I believe that 10% of current output is probably high, moreso in some industries than others.

That is the primary advantage and, I submit, ought to be the primary goal. The chief advantage of producing things on a small scale is that the combined total production of a large number of small producers will be a fraction of the combined total production of a small number of mass-producers. Volume is driven by process for the latter but by real demand for the former. This is a necessary consequence of the systems and methods involved, and the difference in volume is very great. The gains to be had hereby are far greater than any incremental gains in efficiency, especially if the latter undermine the effective viability of the former. The ideas of EF Schumacher and others around less-productive economies deserve wider acceptance.

Efficiency per se is not necessary for sustainability. I know that it has become normal to regard efficiency and sustainability to be for all intents and purposes synonymous, but this is simply untrue. A system can be quite sustainable at c.1975 European levels of automotive efficiency, for instance. The key lies in the need-structures: if one needs few miles one can afford to invest relatively many gallons in them.

So, it is more important in an engine project that the process of manufacture is forgiving enough to survive a lack of central control under small-volume conditions than extreme or even vaguely remarkable levels of efficiency.

I understand your point about bearings: but I think the same analysis needs to be repeated for each component. I don't think making bearings at a small scale is limited by cost as much as by lack of range breadth. It makes sense to look for bearings from someone who offers a wide range, but that does not imply that the manufacturing system structure is other than a repetition of parallel processes, each of which might in theory translate to a mom-and-pop scale. Putting enough of them together to constitute a viable product range might be a case for a worker co-operative type of business model.

But that is part of it: we are really talking about makers of diesel pumps, say, rather than makers of engines. Given the right sorts of communities one could rely on the social and economic relationships necessary for pump makers, founders, machinists, and assemblers to collaborate to their various benefit to develop spontaneously. There ought really to be no need to micro-manage. If other innovators come up with alternatives to any part it should not be regarded as a problem but, on the contrary, proof that the socio-industrial structure is working.

The question is thus not if anyone should need this engine, but if anyone should want one. And that raises the question, how is this design - effectively a suite of interface specifications that people are invited to agree on, in whole or in any ad-hoc part - any better than any design of engine that has by now spontaneously developed that characteristic, like the small-block Chevy, or the BMC A-series, air-cooled VW, Harley-Davidson V-twin, etc.? As I say, it's a worthwhile project, but it is alternative to such engines. I submit that it would gain better acceptance if it overcomes perennial problems with established designs and offers true development versatility and definite ease of manufacture at a small scale.

I shall not go into the legal aspects except to state that current legislation is structured to support the established industry and "roundaboutness", despite noble-sounding intentions. In fact it is my belief that much "environment and safety" legislation has been actively manipulated by industry for decades in order to cultivate a type of market most advantageous to it. Therefore pending sanity an element of conflict, and a consequent vaguely black-market approach from the small manufacturer, will remain.

I think what's missing from the perspective you bring up is time.

If you are going to be buying complex components then where do you draw the line between X number of high-tech components needed to finish the project and just buying the whole project off the shelf?

A few comments:

Eukreign's point about time makes sense, but it is also related to the distribution of tasks among all the people involved in the manufacture of a product - which I think we must remember will not be constant in the sort of scenario I for one envisage, which is hopefully not all that different from what others have in mind. Some might contribute smaller, more intense processes than others, and might consequently be of a type of which there are fewer. The important thing is that the range of technological capabilities in a community is cultivated, not that any given operator possess the entire capability to produce any given product, nor even that the tasks involved be divided up even approximately equally. As a practical example, it is fine if a community develops only a handful of precision machine-shop operations for a much larger number of general fabricators. The capabilities of the average general fabricator needn't place an absolute limit on a design, but may influence the mix and balance of processes the design calls for.

I'd like to emphasize the cultivation of capabilities, because things one needs to buy in now might be locally manufacturable in future. And this might not happen all at once, which is why my prior mention of engines, for instance, that have developed a vernacular character over time is important. Local alternative components increase until local alternative assemblies are possible. In this the choice of a base design is important, not only (or even primarily) for its suitability to the purpose but for its amenability to the desired manufacturing basis. In this a pre-existing vernacular acceptance is an immense advantage. If I were to try my hand at the farm truck (I do, in fact, have some thoughts) I would design it to be powerable by a small-block Chevy or a Cummins 4BT. If any engine is ever likely to be buildable entirely out of diverse obscure mom-and-pop parts, the SBC must be it.

It should be clear that the scenario we are proposing would necessarily blur the boundaries between manufacture, modification, and maintenance. So, too, would the "thereness" we've come unconsciously to ascribe to manufactured goods now tend to give way to a more organic awareness of the processes involved. The age-long cultural distinction between the natural and the artificial has recently become complicated by the addition of a third category: the corporate-made, i.e. goods whose processes of manufacture are so inaccessible that they might as well be products of nature or Providence. To religious, social, and ecological sensibilities alike there is an element of Babelian hubris at work here. Thus we regard modern manufactured good as being "there" in the same way as trees are there - though lacking, to me at least, the underlying sense of goodness. The old artificial, far from being an inferior category, is by comparison much more to be desired, for the mark of my own hand - absent from the corporate-made - only most often much better.

This impacts also on the concept of type, i.e. the Universals Question of our age, in which individual exemplars of a product are intuitively considered physical reiterations of a single reality, much as the medieval radical realists tended to see the world. Except that we want our radical reality to be enforceable through such things as type approval, which would seek out and destroy the rogue or orphan exemplar. I think the idea of pattern as developed by such as Christopher Alexander may be useful here. I should hope that we may see the type give way to the pattern (though, strangely, the terms are reversed from engineering usage, which gives the greater conformity to the latter). This does translate to a design problem that is simultaneously easier and tougher.

My apologies for this tangent into the philosophical.

Matt, I think you need to clarify exactly what you mean by "infrastructure", for both our and your own understanding. I think I have a fair idea of what you mean, and it's something for which I myself am forever trying to find a proper term; but your use of the word struck me as just ever so slightly off the normal conventional sense of roads, bulk services, etc. Capital goods? Means of production? Technological investment? It's all and none of those things, but it is something real and it wants a finger put on it.

But as regards channels, angles, and tubes, the latter is only a weld more elaborate than the others and need not really require a difference in type of operation - so long as they are cold-formed. Hot rolling has certain dimensional requirements that militate against a smaller operation, but cold rolling can be done on the back of a truck, up to a point (i.e. wall thickness). The balance to consider is versatility v. elaborateness of process: the more generic and versatile a thing is, the more its manufacture on a larger scale is justified. Square tube can go into a vast and wholly open-ended range of things, and moreover perform a huge variety of functions in them. It is simple to describe in terms of a very small number of dimensions and parameters, and it is intended to be cut to length. Having known types of square tube available does not hinder but helps the manufacture of more complex things in very small numbers.