Currently, pycovjson dumps out a regular numeric axis as a big array of axis values. CovJSON has a compact notation using "start", "stop", "num" properties. pycovjson should detect if an axis is regular and then use that notation to reduce data volume.