Position nominals - Datum reference or current alignemnt

Position of a width will be reported in a single axis only.

Can't say that I understand clearly what you're asking, but maybe that will help.

Position of the pin should be reported with 2 axes.

Position of a width will be reported in a single axis only.

Yes, drawing shows 2 positions,separate for thickness and width.
So with this we are correct, only one axis.

My second part of question is form of reporting.
in pc-dmis we can choose datum reference in advanced tab, or current alignment.

Current alignment Z=0 (which we measure in)_
Datum reference Z=30 (example, can't remember excaclty)

This changes results quite badly.

seeker, What does your Feature Control Frame look like on the drawing? Where are those datums supposed to be located? I almost always use the Datum Reference instead of current alignment, unless I'm doing something silly.

Sorry can't share more

seeker, without seeing the rest, I can only assume that the original designer made a mistake by including all of the datum references. But first, make sure you have the datums properly defined in your model. Then build out the True Position dimension referencing the Datum References rather than current alignment. Sorry I can't be of more help. Maybe you could create a sketch of the part, without the other dims, or modified dims to help show the datum structure.

Your print you shared depicts slots, not pins.
I think your conundrum is the pins can freely move within those slots and you don't know how to adapt your rotational "play" between the two slots to make the positional control a straight linear distance from the first datum(D) pin?

if that's the case...
Measure D. origin to it.
if your second pin is free to drift anywhere within that second slot, i would add a "find the pin" kind of point, where you take a vector point parallel to the length of the slot with a really high prehit/retract and CHECK distance. Origin to the found pin point + the pin's theo radius
consturct/point
findpin.tx, findpin.ty + (or minus) PINRADIUS, findpin.tz, findpin.ti, findpin.tj, findpin.tk
findpin.x, findpin.y + (or minus) PINRADIUS, findpin.z, findpin.i, findpin.j, findpin.k

origin to this constructed point, measure the pin as an autocircle at xy 0,0
now, recall your alignment back to D.
constuct a line between pin and D
rotate to that line
report position, with your 2mm BASIC distance hard-set.

I always report in the current alignment (except for a very few circumstances).

Maybe neil.challinor can chime in with the procedure of the GeoTol alignment - which datum axis is which when reporting according to datum reference?

Primary datum = X?
Secondary = Y?
Tertiary = Z?