I am sincerely sorry for my harshness in the discussion of figured bass. It came from concerns about the typesetting of more modern music. I should consider it natural that many would be most concerned with old music, if only because it is all pd, and pd goes to the mudela site, which is a worthy goal. I was a pig to say what I said. There is plenty of ignorance to go around, and I confess to having more than a fair share of it. The stems extend to the middle of the head. They should extend to the optical (in an artist's sense) meeting of the stem and the long axis of the notehead. The extra length is wrong, looks bad, looks *very* bad at large sizes, and leads to false collisions in the case of the unison, which should be a different issue, but unfortunately is not. Apologies if this is fixed already. I complained about the wrong thing last time. Nobody picked up on it. Too bad. Sorry. Please offset all stems in all cases by 0.25 whether or not they collide and regardless of direction with one exception. 2nds with opposite stems should be shifted in the *opposite direction* from that of the others. You are probably not the first to misunderstand that. Try to be the last. The amount of negative shifting depends on the design of the noteheads, since they probably would butt together. Again, I plead guilty to not enough study of the output to get full understanding of what was bugging the hell out of me before. Again I am amazed to be the one to point this error out. If this error resulted from an example you saw rather than statements in a book you misinterpreted you should never trust that engraver again. Force horizontal note shifting should have resulted in a better visual balance vertically. That was its purpose. It does the exact opposite in the case of the 2nd. That is not right. I like the horizontal separation in the case of the unison, although it results from a mistake. It could remove ambiguity if there are seconds or other unisons on either stem. It also just plain looks good. There is no vertical visual balance to improve. Please put it back in after you remove it by fixing the stems. :-) Just because there existed a policy of horizontal shifting for stem collisions by half a notehead is no reason to do it that much. Remember that publishers are allowing for inaccuracy in drafting, which lilypond should not have to do. They wanted a target or average, you want a minimum. 0.25 is plenty, more elegant, perfectly legible, simpler, and often seen. It tends to better horizontal spacing. Allowing no fault tolerance is especially good, since lilypond engraves perfectly ;-) and won't let the smaller space between the lines appear to lack uniformity, which lack often occurred in engraving and typesetting. Advantages: 6 so far. Disadvantages: none. Summary: 1 Shorten stems at the head end. 2 Get rid of forceHorizontalShift and just fix HorizontalNoteShift. 3 Where there are two stems or more, shift the topmost stemdown to the right 0.25 if it is a 3rd or unison from nearest stemup, otherwise from topmost stemup, if it is a 2nd shift to the left instead. 4 There must be no shifting with a silent note, since that would result in a stem connected to nothing. Silent notes (headrests?) are badly needed. You should have them by now. Thank you, bless you, work hard! .daveA
