Steve Gardham wrote: if I'd tried to include everything in the DT or Roud Index I'd be pushing up daisies before it got finished.

And, as a secondary problem, it's effectively impossible to keep track of all that stuff. The single largest component of my correspondence with Steve Roud is places where I've noticed that two songs should be combined under one number. Tracking everything that has ever been published as a traditional folk song is something only an extreme savant could remember (and an extreme savant would probably have trouble identifying them, because of the limits of savant-hood).

Like Steve Roud, I try to include everything, but I don't obsess about finding every fact about every song. Instead, I tend to try to find what is easily found -- and pick a handful of songs for deep analysis. That's not the same as what Steve Gardham is doing with his Earliest Dates (he's chasing versions; I'm chasing background information), but in both cases, it's making a near-infinite task finite.