On your first point, your friend is absolutely correct for the majority of VOR holds. Where the hold is based over the VOR, the joins are exactly the same as for a hold over an NDB, with the three standard entry sectors. Where VOR hold entries can differ is either where the hold is at the intersection of two VOR radials or where it's defined by a radial and a DME distance. In those cases, the entries are limited to the inbound track, a published track or by radar vectoring. If you look in PANOPS Doc 8168 Vol I, there's a pretty good representation of the different entries from different directions. In reality I think the hold entry options are best summed up as "do something sensible". VOR holds based on intersections or radial/DME are relatively rare and usually the expected entry is blindingly obvious - for example when you're approaching along the inbound track.
On the second point, the abeam position will only be at 90 degrees to the aircraft either in nil wind or where the wind is blowing parallel to the hold axis. If there's any wind component across the hold, your outbound track will have wind correction added to it, which can be quite a significant amount on the outbound leg even if the inbound wind correction is small. With, say, a not uncommon 30 degree wind correction angle, the NDB will be at 60 degrees to the aircraft axis, and the ADF will suffer from dip errors with any bank on.