...that's not an easy question! Here goes:
Accepted wisdom (and Certification):
An aircraft may be considered a bunch of pressure fields; and as they are sufficiently intense to hold an aeroplane up, it is accepted that it is virtually impossible to produce an ASI system that is both suitable for everyday use, and very accurate in all possible operating conditions. The atmosphere keeps changing, too...
Therefore, Certification requires that the airspeed indicator system be calibrated, and the corrections made available to the pilot. However, most people don't like reading correction cards halfway down finals, so gradually design standards have reduced the original wild freedom AISs once had.
The calibration system involves getting the static system out of the aeroplane's pressure field, by towing it on a drogue ~ 30m below and ~100m behind (less for lighties) the aeroplane (there are pictures of jet airliners with plastic funnels on hoses hanging off them!). The pitot ends up as a pivotting (self-aligning) type on a boom several chords in front of the wingtip, or if flutter is an issue, as far fowards of the LE as possible.
Practical Systems:
1) Static:
The laminar sub-layer of the boundary layer, in air, is always very close to static pressure - if no extra pressure field is superimposed. The innermost edge of a turbulent boundary layer - outside the laminar sub-layer - is also very close to static, if unaffected by external fields. As the boundary layer gets thicker moving aft along the fuselage, and the main causes of external pressure fields - the propellor and wing - are normally towards the front of the fuselage, putting the static port(s) as far back as can be - without the tail affecting things - is good practise.
Unhappily, a sideslip will cause a pressure difference from side to side of a fuselage; and it's NOT symmetrical - "T"-ing ports on both sides reduces the error to ~1/2, it does not eliminate error. Furthermore, putting the ports on little raised discs of aluminium because it's cheaper than a flush port, risks raising the ports into that part of the boundary layer containing fluctuating pressures (the fluctuation is too rapid for the instruments to detect, but the mean pressure will vary with speed).
Symmetrical flush ports 3/4 of the way back on a roughly circular glider boom rear fuselage work very well.
However, if a superimposed pressure field affects both the static and pitot pressure ports the same amount, no differential - or error - is shown on the ASI (due to superimposed pressure). So, mounting the pitot & static ports very close together gets around the static problem.
2) Pitot:
Most forms of pitot - whether bullet-nosed, or conical, or just bits of tube with the end hacked off - give quite accurate results at up to ~20 degrees misalignment. All airfoils (unstalled) have two zones where the positive pressure of the stagnation point becomes the negative pressure of the accelerated flow (both above and below, generally); passing through zero. It has therefore been standard practice to poke the pitot out of the LE, 0.1~0.2 chords fwd of the LE and about level with the foward extremity of the airfoil, and tweak it until the AIS meets the design standard (needless to say, the slipstream is to be avoided!).
Someone who had studies their NACA reports, noticed that there is a point about 20%~30% back along the underside of most airfoils (unstalled) where the local static pressure does not change much with AoA; and also recollected that the airflow adjacent to an airfoil, is essentially parallel to the nearest surface (except right up front). They therefore stuck a pitot port and a static port on a small pylon under the wing (~0.05 chords high), and filed a bevel off the trailing corner until the pitot-static difference represented the free airspeed. Look under a Piper Cherokee wing. The static error is not an issue; crossflow is not an issue (the wing acts as a flow straightener); sideslip is not a big issue; and few people bang their heads on it.
That'd be my first choice. A combined pitot-static head, as far outboard along the wing (but not near the ends of any lift struts) as practicable, and ~0.15 chords ahead of the LE would be my second. Note that putting a shroud on a pitot gives up to ~30 degrees of misalignment without significant error, although the shroud supports have to be faired and well behing the dynamic (pitot) port.