The Art of Electronics, combined with the Student Workbook are excellent books to teach yourself electronics.
I guess today you'd use electronic simulations on a computer rather than real components on bread-boards with real test equipment. Perhaps some HNer is involved with this type of simulation software? I might be worth creating "labs" that can be used while reading tAoE?
I think the best analogy I can give for doing electronic labs online is the way cisco certification tests were provided in a very limited simulation presumably using regex's to see if you get it "right" (at least this is how it was a decade ago) vs getting time at an actual router command line. Its a different experience to do something real vs a simulation with pre-determined conclusions and situations.
I'm struggling to think of a programming analogy, comparing getting a real live REPL to getting some kind of not-REPL training environment. I don't think there really is anything that bad in all of program writing.
What simulation is really good at is optimization. Its a useful skill but not the only one. Maybe a good programming analogy would be ripping all the "write a program" assignments out of a CS curriculum and replacing them all with profiler exercises. So rather than writing your own bubblesort and quicksort, you'd just run a profiler on someone elses sorting libraries and compare the numerical results to get the expected result from the book.
A really good car analogy is I'm old enough that when we did Drivers Ed we had simulation where driving scenes were projected on a screen and we optimistically pretended to drive a car, vs the behind the wheel section of drivers ed where we actually drove a real car around. Its kind of useful, kind of, but I don't think you can really learn to drive a car by watching carefully crafted movies of someone else driving.
I studied the Cisco networking materials 5-10 years ago but never actually got round to taking the tests. Dynamips was a router hardware emulator for the 1700, 2600, 3600, 3700, and 7200 hardware platforms that booted up actual Cisco IOS images.
With GNS3, it was possible to set up complex network topologies of dynamips routers that would have cost thousands of dollars in real hardware without losing any of the realism as these were essentially virtualised routers.
I haven't kept up with the progress here but Cisco certification may not be the best example.
I guess today you'd use electronic simulations on a computer rather than real components on bread-boards with real test equipment. Perhaps some HNer is involved with this type of simulation software? I might be worth creating "labs" that can be used while reading tAoE?