As a first test balloon for mapping with compass and distance, I choose two locations: Piazza della Riforma and Via Nassa, both in Lugano's old town.
But before I could start, I needed to calibrate the step counter. For that one needs a known distance. Conveniently, the railways station is on the way and, even more conveniently, the length of a train is written on its side. So I found out that my average step is 85 cm long. Verification along the same train showed that it was off by less than a meter along the 75 m long train. Good enough.
Downtown I started with Piazza Riforma. I GPSed a reference point in the wide open, sketched the piazza and then took directions and distances. Afterwards, on the computer, it turned out that some of the directions are as much as 30 degrees off. Whether this is me not yet knowing the compass well enough or this is just the normal deviation of a compass, needs to be determined.
Via Nassa went altogether a lot better. There already is a trace of it in the map. My newly determined course is extremely close to that one, except in some parts where the old one was clearly wrong. Oddly enough, on the last bit very close to Piazza Riforma the direction was off by about 15 degrees yet again.
So, next thing is to try the piazza again, maybe with a couple of extra GPS-based reference points and some more care when taking directions.
Discussion
Comment from chillly on 9 November 2008 at 13:30
I would expect a compass to be accurate to a single degree in most circumstances. They can be affected by iron nearby. If my compass was even a few degrees out I couldn't safely use it to navigate across moorland in a fog, which I have done may times. Using a consumer GPS as a starting reference point for surveying introduces a lot of error when mapping a small space like a piazza.
Comment from partim on 14 November 2008 at 20:34
I am guessing it is mostly me not being familiar with a tool just yet. Didn't really expect to get perfect results at the first try (My first GPS trace was completely useless).
Regarding the reference point: Even if it is extremely off, the shape of the piazza should be roughly right, anyhow. Which it wasn't. I think that it was mostly a matter of small errors adding up. Will try again as soon as I have time.