High resolution line imaging notes

1. Don't use regular CLEAN. It's going to work very poorly.
2. All of the archival low spatial resolution data are going to be at higher resolution than the high spatial resolution data. Even so, it seems best (based on some experimenting with the TW Hya data) to leave everything at native spectral resolution before doing line imaging.
3. Multi-scale CLEAN is ostensibly able to tolerate higher gain values (0.5-1.0) compared to regular CLEAN (usually recommended at ~0.1). In practice, it seems that a gain value of 0.1 still leads to visibly better results for cleaning lines with multi-scale CLEAN. It apparently matters less for continuum emission imaging.
4. Multi-scale CLEAN seems to be able to handle CLEANing down to lower flux thresholds compared to regular CLEAN (e.g., see Rich et al. 2008)
5. cvel does not behave nicely when you've got multiple SPWs in a measurement set.