Some 50 ohms inputs support AC coupling but most do not. 50 ohm circuits often expect and require a DC termination to ground. It is easy enough to add a DC block if AC coupling is required.
The Tektronix 7A19 vertical amplifier is an example of what is required to safely support AC coupling with a 50 ohm termination.
1. Why they didnt put option to switch to AC coupling on 50 ohm?
It adds complication and most users do not require it. If needed, a 50 ohm DC block may be added externally.
2. Isnt AC coupling just a simple high pass filter? Cant it be simply made in DIY without signal degradation?
Sure, but large high frequency capacitors are not trivial and make maintaining a 50 ohm environment difficult.
3. Dont scopes have inbuild hardware offset function as default? Can it be used in place of power rail probe offset?
The internal 50 ohm DC termination comes before any offset capability (1) so the safe input voltage range is not increased. Measuring voltages above about 5 volts will require AC coupling or an offset probe.
Typically DSOs have a vertical position control, but all that does is change where the trace is drawn onscreen. If the signal coming in is outside the current range it will still clip. If you want to use the full precision of the scope between, say, 10-30V (using the -10V to +10V range), you would need to add an offset at the probe.
In an analog scope, the vertical position was in some sense changed by adding an offset voltage, but that happened post-attenuation, so it still wouldn't be a solution to the problem of using the whole range.
I don't know about more modern DSO's but my LeCroy's apply a true offset to the input.
In fact sometimes I would like a control just to change the position (on the LeCroy you can cheat and use the maths traces which do have a "position" setting).
Helius is wrong about this; only the cheapest DSOs operate like he describes. Most DSOs which only have a position control including modern ones implement it as a combined position and offset control just like analog oscilloscopes do. The combined position and offset signal is injected after the input attenuators and buffer stage and before the final attenuator or gain control stages. This is reflected in the specifications which show different position/offset ranges for only some attenuator settings.
For example, a Rigol DS1000Z series has a position/offset range of +/-2 volts at 1mV/div and 8 vertical divisions (2) so its position/offset range exceeds the ADC resolution by up to 500 times because of the gain after the point where the position/offset signal is inserted. It has two separate offset ranges because it has one switchable attenuator before the point where the position/offset signal is injected.
(1) This does not need to be the case but I do not know of any counter examples. The 50 ohm termination could be offset just like how a low impedance offset probe does it but this would be an invitation to lots of blown out terminations and front end circuits.
(2) As far as I know, Rigol shows the entire ADC range on their 8 vertical division display. Other DSOs often clip at 10 or 12 divisions while showing 8 divisions which has some processing advantages.