Even in a ballpoint pen, there a good 5-10 dies needed to make the body and cartridge holder, pretty exotic processes to make the tip, a non-trivial pad-printing for color, and many other post-processing needs. Yes, we could reengineer products to not need this, but you're going to see increased costs and decreased performance - I.e. No one will buy it...
I think I adequately addressed the "increased costs and decreased performance, so no one will buy it" objection in my original comment. :)
That's at least party due to the different design requirements and constraints that mass-produced for-profit engineers work under. That lawnmower wheel screw was specified not purely on the engineering requirements to "hold the wheel on", but on various other important-to-the-company-designing-them-for-sale requirements, probably including things like: use the cheapest fastener that we already keep in stock that'll do the job, use a fastener that will reliably last the length of the warranty period at the lowest cost, use a fastener that's compatible with out existing automated manufacturing tools. On top of that there will have been other decisions already made like "use an existing wheel assembly" and "ensure the mowers packaged size and weight fit in with existing supply chain logistics".
Once designers are freed from existing manufacturing industry constraints, I like to hope that one day we'll download a push-mower design from somewhere like Google Sketchup, where the designers were perfectly happy to "over engineer" critical components and where the design requirements might include ideas like "should last long enough to pass on to your grandchildren", and "will still be able to be repaired at the side of the road with a flatblade screwdriver and a rock".
I'm actively fighting for open source hardware and standards, but a big part of this is finding off the shelf components - look at Http://octopart.com for a good example of what this looks like.
Oh, very much not.
I'm more commenting on unnecessary overspecialisation instead.
Unnecessary for non-commercial applications anyway - I understand why Apple chooses to put difficult-to-use fasteners in their products, if for no other reason than to minimise fraudulent warrant claims, but there's a whole different set of motivations if designers are going to use the equivalent of "open source" for digitally reproducible home-fab ready "object" files.
Specialized raw materials are a problem, but I have the impression that a remarkably large fraction of everyday items could be cut out of a sort of general-use steel or brass with only a marginal increase in size. There are lots of different steels and brasses (and of course a huge variety of other materials) but you don't really need them for most things.