Jim --
I own a 100 meter steel tape but I can't find it. If I ever do, it's yours. It's an "add-meter" style -- the body of the tape is marked only every 5 meters (brass soldered to the steel). Then on one end is a fully-graduated (5 meters by mm) segment. To use it the person at one end holds one of the 5m marks over the point, and the other person uses the graduated end. Then somebody gets to add the two, thus the name. In fact the tape is 105 meters long. It's very narrow, but durable. I used it while land surveying in the Solomon Islands rain forest and it held up nicely for years.
On the matter of GPS units. I have several. From the point of view of a surveyor, every one I have (and every one mentioned here) is a consumer/hobby unit. But that's appropriate! If you want to get into GPS as a land surveyor would, then you're into "differential GPS" at a minimum, and your unit is going to cost you a lot more than my Garmin 60cs cost me.
One problem I haven't seen mentioned in this thread has to do with hills (or even undulations). A bicycle with a Jones counter is always following the road surface. A distance measured via EDM (for example) does not. If the road is perfectly even, then the EDM measurement should be the same as the calibrated bicycle measurement. But if the road surface rises and falls, it won't be -- the calibrated bicycle measurement will be longer. I used to use a calibration course that was measured by EDM, but the road undulated. I used it anyway, having figured out that the different was not great.
What this has to do with GPS-measured courses is that a GPS unit which does not or cannot compensate for hills will give a short course, other things being equal. Imagine starting on the flat, running up a hill, down the other side. The GPS unit is really dealing with latitude/longitude pairs, and it calculates distance by calculating where it is in relation to where it's been.
A Garmin rep told me recently that their units attempt to compensate for this by factoring elevation changes as best they can, to estimate travel along the actual route. But -- as anybody with a GPS should know -- ordinary units do elevation poorly. If you doubt this, take your unit to the seashore or to a point of known elevation. My 60cs has a barometer and thus is better than most, but it's still not very good. Why, you ask? For a GPS unit to do altitude as well as it does lat/long, it would need a sat on the other side of the earth, and that's impossible.
I like my units (including a Garmin 205) and use them all the time. But they don't come close to a calibrated bicycle. I stopped measuring courses for certification ten or fifteen years ago. If I started again I'd always have a GPS unit with me because it protects from blunders.
And as for starting races at the correct points -- when I was in the timing business I routinely saw race directors not able to remember (or find) the points, or shifting them for one reason or another. It was rare for the director to have the certification map on hand. All this is bad behavior but it happens.