It turns out to be easy to compute a good estimate of dew point from humidity and temperature. Please compute and report this for humidity-and-temperature sensors that do not already report dew point. That would give use many more sources of dew point.

It is only an approximation, but it is almost certainly exactly what our existing dew point sensors are doing, since the alternative is extremely complicated: induce condensation by controlling the temperature of a mirror.

There are actually several formulas available, but the most common is the Magnus formula, and it has the virtue that it only uses humidity and temperature.

ts_watcher has the following, which may make this easier for you:

• An implementation in DewPointFromHumidityWrapper in rules/dew_point_depression.py
• A unit test in tests/rules/test_dew_point_from_humidity_wrapper.py
• A pdf describing the formula (in the doc directory). This is from Sensirion. They manufacture of a different dew point sensor than the one we use, but it was the nicest reference I found, and I could not find the information for how our sensors compute dew point. Wikipedia has the same formula (using different notation), as do many other web sites. So you have your pick of references.