Power can be calculated as:
Power (watts) = Voltage x Current (the "basic formula").
Ohm's Law describes the relationship between current = Voltage/ Resistance (resistance and impedance is
"the resistance to current flow".)
From those you can derive:
Power = Current Squared x Resistance
and Power = Voltage Squared / Resistance. (This formula is often handy because it's easier to measure, or know, voltage than current.)
So according to my handy-dandy spreadsheet,
140W into 8-Ohms is about 4.2 Amps (and about 33.5V).
Matching Impedance:
Is it crucial the amplifier's output impedance match the speaker's impedance (8 ohms in this case) ?
Wrong! 
The amplifier has a
rated minimum
load impedance. i.e. An amplifier rated for 4 Ohms can be used with 8-Ohms speakers but an amplifier rated for 8-Ohms can't be safely used with 4-Ohm speakers. Amplifiers are "voltage sources" which means they will put-out the same voltage at 4 Ohms, 8-Ohms or with no speaker connected (infinite impedance). If the amp isn't rated for 4-Ohms, you'll "pull" excess current, the voltage may drop and "bad things" can happen.
The amplifier's actual output impedance is usually a fraction of an Ohm. You won't find a spec for it but sometimes you'll see a spec for
damping factor which is the ratio of load impedance to the amplifier's output impedance.
Power Rating:
Does the amplifier's power rating need to at least be equal to, but ideally greater than, the speaker's continuous power rating ?
Power Spikes:
Speakers can experience sudden power spikes during music playback, so is it's a good idea to choose an amplifier with a bit of headroom ?
Rule of Thumb:
Is it a common rule of thumb to choose an amplifier that can deliver 1.5 to 2 times the speaker's continuous power rating ?
No. 140W Speakers are designed to be safe with a 140W amplifier that's hitting 140W on the program peaks (with a much lower average power level). It's "statistical" and
approximate because it's the short-term average power that burns-up speakers and the peak-to-average depends on the program's (unknown and variable) dynamics.
You can burn-out a speaker with full-power constant test tones, especially the tweeter which can't take as much power as the woofer.
You can burn-out a speaker using an amplifier with more power than the speaker's rating (if you actually "use" the available power).
You can burn-out a speaker by pushing an amplifier with a "matched" rating into clipping/distortion. When it's clipping, the peaks are limited but the average power continues to go-up as you turn-up the volume and the clipping generates harmonics some of which go to the tweeter. *
Example:
Is it any detailed improvement if your speakers have a 140-watt continuous power rating to have an amplifier rated for 280 watts (2 x 140) instead of 50 watts per channel for medium volume level ?
You don't gain anything from headroom unless you use the extra power. And if you're using it, it's no longer headroom!
I`m aware that higher current adds more precise control
Nonsense. A higher damping factor can give you more "control" but with solid state amplifiers it's not an issue. Some tube amps have a poor damping factor.
My Harman Kardon AVR 140 is only putting out 50 watts per channel in stereo mode, probably good I`ve always lived in Apt or suites and have never turned them up to their full potential ?
As above, extra-unused headroom gives you nothing.
AVR 140 is also Rated 25 amps if that matters...
That would be 5000 Watts! Kind-of a meaningless spec... SOME "extra current" capability CAN be helpful because speaker impedance isn't "flat" across the frequency range. It's not unusual for the impedance of an "8-Ohm" to drop to 4 or 6-Ohms at certain frequencies, and since the lower impedance will try to "pull" more current, if the amp can't supply the current, the voltage will drop and the amp will clip.
* There is a popular myth that a higher power amp is safer than a lower-power clipped amp. Both are "dangerous" and the higher-power amp is worse.
P.S.
for full resolution maximum detail ?
Resolution mostly relates to digital. I suppose noise (hum, hiss, or whine in the background) reduces analog resolution but we usually just talk about noise. BTW - If you listen to an 8-bit file the loss of resolution is heard as "quantization noise".
Detail isn't really defined either. It makes me think if high frequencies... Boosted highs = more detail, rolled-off highs = less detail... BUT surprising to me, Dan Clark (headphone manufacturer) says certain distortion is often described as "more detailed".
