I wish I had an oscilloscope. I wish it BAD. Why? Because just having a good multimeter really isn't enough to understand what's happening inside the circuits I'm building. Not even close.
It's not enough to know voltage between two points on the circuit, though that IS useful sometimes. And it's difficulty to measure current without disconnecting and reconnecting things. Those are both tools I have with my multimeter from college and I do use them.
But to actually see, say, what's happening to voltage along a line when you have a 1k ohm resistor charging a 220uF cap, versus a 10k ohm resistor charging a 220uF cap, you really need continuous voltage monitoring over time. The multimeter can tell me roughly what the average is, and I can see if it's fluctuating at all, but with a scope you can see actually what's occurring over time, and that's hella useful.
So I built a couple simple PC-based scopes. I wanted more detail, and I didn't want to spend the hundreds of dollars that it takes to get even a half-decent entry level standalone oscilloscope. I think I can probably find a decent 20 year old Tektronix one for like $350 on craigslist or eBay, but it's not even a storage scope, which is something I'd want to have for digital circuit work. Good new ones are in the multiple thousands of dollars, it's all very discouraging.
But a little searching the web turned up a couple of somewhat viable options. They don't sample at NEARLY the frequency you'd need for "serious" scope stuff (like video) though.
A soundcard based scope:
A parallel port scope:
Both of which cost actual money money for the software (which, I'm finding, is really important) but the barrier of entry is low. With the soundcard scope, it's just a couple of resistors and some diodes to make a small circuit to protect the LINE-IN input on the sound card from over-voltage. With the parallel port scope, a couple of cheap ($2?) 8 bit AD converter chips were necessary, but the circuit itself was just a voltage regulator, some resistors and those chips.
I built both of them. Why? Because the soundcard based scope has a fatal flaw. While it's great for analog stuff, there are some capacitors on the line-in on the sound card and they normalize the voltage. If you hook the probes up to a AAA battery, you'll see the 1.5v trace, but then it immediately tapers down to zero. This makes it unusable for any kind of digital work since 1's don't stay 1's for long (at least visually.)
So I built the parallel scope for digital work. It behaves exactly as you'd expect an oscilloscope to behave when you hook up a 1.5v AAA battery to the test leads---you see the 1.5v trace. I hope that this is useful for digital work, even though the scope's maximum sample frequency is 4kHz compared to 44kHz for the soundcard scope and 20Mhz for a very bottom run entry level professional scope. I guess time will tell.