The better solution would be to teach people how to use the parts
If your attitude is that you know best what other people should do, you won't get very far. They don't want to plow through a book, they want to build something that does something. Who are you to tell them that they can't?
You say that getting them interested with a kit is useless because they will bail as soon as they hit some math. Well, they're certainly no worse off than if they hadn't built the kit, which is the alternative.
> The problem is no-one knows how to use the parts unless they are assembled into kits.
Huh? If anything, in the good old days, there were WAY more kits than there are today. Besides that, hobby electronics has a long history of doing stuff like "Here's a schematic, here's a BOM, and here's a paragraph on the 'theory of operation' that assumes a solid understanding of electronics and a bunch of domain-specific stuff. Good luck!" -- Pretty much the same as it is today, except we have resources like http://electronics.stackexchange.com/ to answer our questions. Wish that stuff was around when I was younger.
Kits just save you the trouble of making your own PCBs and hunting down parts on Digikey.
I think there are more kits now. There used to be kits for oscilloscopes, test equipment etc. Now these are black boxes, some of which are a joke.
We had something better than stack exchange - we got an electronics book and a maths book out of the local library and studied it, understood it and developed mental models, rather than expecting a canned answer to plug in somewhere.
And yes I have scars from ferric chloride accidents.
So everything studied is instantly understood? People have no need to ask questions? Have mentors?
Sure, some people are looking to plug-and-play answers, but that will always be the case. You can't solve that by telling the kids to 'get off your lawn.'
"The problem is no-one knows how to use the parts unless they are assembled into kits."
You're making a big assumption. Not everybody has the time to design, source and test a complex system from the component level. (Nor do programmers drop down to the machine code level for every project.) Many kit builders have the background, but still must juggle other priorities in their life.
Ideally folks should learn what's going on at the component level. With the free resources available online, that's easier than in the past. But no one should be forced to go through Boylestad or NEET just to play with a LED.
Capacity for understanding isn't a limited pool. If people get excited programming micro-controllers let them. They really don't need to learn to build such a thing.
Should all programmers learn assembly? Sure. Should all programmers be forced to learn assembly before anything else? No.
"The problem is no-one knows how to use the parts unless they are assembled into kits... that's my problem."
Funny, I've designed and built from scratch all sorts of advanced analog and mixed-signal circuits using my "secret" EE knowledge. I've also built kits from Adafruit. I had a great time doing both, and do not consider them mutually exclusive.
I don't think the ignorance of _other_ people is your problem.
The problem is no-one knows how to use the parts unless they are assembled into kits.
The whole maker movement solves this by making kits, therefore raising the bedrock abstraction for knowledge way too high.
The better solution would be to teach people how to use the parts, which is my problem.