The IFAB said that they changed the location of the restart because this would make it consistent with all other offences where the restart is taken from where the offence occurred.
[...]
Offside is an offence which is unlike all others in its nature, so I don't see that it should be consistent with other offences in terms of its restart. No other offence involves two separate events that occur distinct from each other in both time and space, in the way that offside does. The second part of it - involvement in active play - is indeed the culmination of the offence but just as the first part (being in an offside position) is not an offence in and of itself, neither is being involved in active play. I think it's a slightly fallacious argument to say that the offence only occurs where the player becomes active
IFAB's big push on the big rewrite was to introduce consistency across the board -- there were already far too many "exceptions to the rule". The offside one? Just another of those. They still haven't eliminated all of the exceptions, but they've made it a LOT easier for new referees to figure stuff out.
In terms of offside being two separate events, it's the same idea as a player receiving the equivalent of a red card (2nd caution or straight red), but advantage being played, and then the player getting involved again. In the case of the offside offence, it's almost like it's coming to fruition, and the offence has occurred at that point.
The biggest issue here (and let's be honest) is that most ARs are still stuck with their old habits of STOPPING at the point of potential offence and not following the play until the PIOP actually becomes involved. As soon as that stops, then it really becomes a no-brainer.
Remember that prior to the re-write, the location of the kick was supposed to be where the PIOP was located when the last touch by a teammate occurred. So... if that player finally interfered with play/opponent/etc near the PA, but was originally in that offside position near the halfway line, strictly speaking, the AR needed to raise the flag at the point of offence, run up the touchline to the original position of the player and THEN indicate the location of kick...
Much easier now -- when the flag goes up, point to location, and hey... the location of the kick is almost always where the ball is (was) at the time of the flag going up.