Hint: there is no such thing as digital. There is always "slew" There is a rise time and fall time to each "pulse".
HOW much redundancy is required for PSK to "work"?
CW puts all the energy on ONE "analog" frequency, and the EAR and BRAIN is better equipped to "analyze" an analog
audio waveform and interpret than letting a "logic level" discriminator decide if it's 1 or 0.
I'm NOT knocking digital. I live in world of both analog and digital, and can tell when each is or isn't "working".
With the old analog cell phone, I could hear through the flutter of a weak connection.
With digital, I can hear the phone "giving up" much too easily, and deciding there's no point in trying.
With analog TV I could "see" some snow, and the desired signal at whatever level.
With digital, I see a message at the top of the screen saying "NO SIGNAL" when I know darn well it's just not
enough of a stream to begin reconstructing the data.
I have no problem imagining a full power digital band, or how well it works, we have the testimony of many who now cannot receive television signals in places where they used to be able to.
Digital is great in a medium which WORKS for the mode, like water in a pipe.
In fiber optics it is absolutely the best, most loss-free and best way to move data.
Say, when are they bringing that fiber optic system around to my neighborhood?

Haw haw.
You can transmit "pulses" of water pressure in a pipe quite nicely, but don't try it in mid-air.
Radio works as simply as wave motion in water when it's analog, and suffers exactly the same interference.
You cannot transmit "pressure" pulses through open air with water with any more sucess than digital in lower frequencies
of RF. Make the wavelength short enough and it starts to look and act like it's in a "pipe", and OK, my computer works on
3G or wifi. But it's not radio then, it's like the behavior of light.
When you can get all the QRN and ionospheric aspects to play nice, I'm ready to migrate radio to digital.
Until then, I'll take my lightning crashes, deep fades, and intermodulation over a blinking blue silent LED any day.
I can make a blue LED blink without all that overhead procesing and in "real-time", too, no 8 second delay.
"ONE" of the complaints? How 'bout the intrusive iboc noise in multipath spots on FM?
1% is too MUCH!
In a spectrum defined specifically for digital, I'm OK with anything you wanna try, and I'll buy one and try it out.
I could even be a big booster for THAT!
When someone tells me that a digital signal can be anywhere near (spectrum-wise) an analog signal, and I won't hear intermodulating and mixing with the analog, I have enough of an engineering background to see through such dishonesty.