The GUND value seems a weird thing to specify on approach plates, because any IFR certified GPS already contains a lookup table (covering the whole area of the basemap provided) containing values for the difference between true MSL and the geoid.
So, the IFR certified GPS will already read the correct altitude. As indeed my KLN94 does, everywhere I have been.
Some handheld units have this data in them too. This, I recall, was one of the differences between the SIRF2 and SIRF3 chips. My yoke-mounted Garmin 496 also reads the correct altitude, within a few feet. But my Emtac bluetooth GPS is 200ft out - same as most cheap GPSs are around the UK.
You want to post this in Usenet in sci.geo.satellite-nav if you want real expert input. I can post it there if you don't know how to access Usenet.
BTW there is nothing "stupid" about using GPS altitude. With the right kit, it is very accurate. Not as accurate as an altimeter set manually to read "310ft" when sitting at an airfield with a published 310ft elevation
but a whole lot more accurate than the same altimeter will be once you are airborne to say 10,000ft, at some odd temperature to boot. With extreme deviations from ISA, the altimeter could easily be 500ft out at 10,000ft. The reason
barometric altitude is used in aviation is for mutual separation of traffic, which is a different thing from absolute altitude accuracy.
Back to that Goodwood plate you reference, it seems particularly daft to litter it with not only the four GUND values, but also lat/long coordinates for the runway thresholds, given that nobody should actually be entering lat/long coordinates manually, ever, for any serious purpose (like flying approaches). A certified database will always be mandatory. I sense some little man at work here, being extremely anally retentive