I found that SODA is rarely used in recent papers but I don't see how others account for the problem that minimum stopover length (as directly estimated by ringing data) can dramatically differ from real stopover.
You're right! Always use real stopover length (RSL) i.e. the period between first capture and the bird leaving the site, rather than a minimum based on ringing data, as you say :)
Guérin, S., Picard, D., Choquet, R., & Besnard, A. (2017). Advances in methods for estimating stopover duration for migratory species using capture‐NDASH‐recapture data. Ecological Applications.