That's because they're converting from the linear to the
side-by-side, and the linear doesn't change by swapping from
one side of the cable.  But a dual row header is already in
side-by-side format, and putting it on the other side of the
board definitely does change things:

    1  a   2
    2  b   1

  Signal a ends up on pin 1 or 2 depending on which side of
the board the header is on.  It works quite well, and .  It
doesn't 'FIX' the dip to header mismatch, it reverses the
signals to correct for the second reversal.
Alan


Richard Martin wrote:
>
> I hope it is clear that changing the 'side' of the board
> or the 'side' of the cable does NOT fix the inversion of
> odd and even numbered pins that results from the unfortunate
> layout of DIP to IDC headers. Do the thought experiment of
> 'sliding,' an IDC connector down to and around the end of the
> cable, doesn't change pin-to-wire mapping.
>
> >
> > Don McKenzie wrote:
> > > As the board is also multi purpose, it's only when you want to use it as
> > > a pod or emulator mode, that the connector needs to be mounted on the
> > > !!!!solder side!!!! of the PCB.