here's your huck finn update for today... *** i reduced the diffs from ~3,000 to ~2,000. so now all of the words/letters are "correct". (the job of standardizing word-forms over the entire book becomes very problematic when dialect words get phonetic spelling.) so now the ~2,000 diffs are on punctuation. based on my memory of punctuation in resolving ~1,000 diffs so far, i'd say that approximately 95% was right in the o.c.r. if that were to hold up, it'd mean we now have roughly 100 o.c.r. punctuation errors. i'm not sure it is worth it to resolve those ~2,000 diffs to fix 100 punctuation errors. (although, to restate the obvious again, that would be much more efficient than checking _all_ the punctuation manually, which is the workflow used over at d.p.) so this experiment confirms my earlier experience that different editions _can_ make comparison less worthwhile to do. (depends, of course, on how many edits were done to the text in the new edition.) it's especially true if you have the option to use an independent digitization of the _same_ edition, even if it'd be some work. as i _do_ have an independent digitization of the same edition, i will change course to execute "the final push" for this book. -bowerbird