English: Shows an excerpt from a german rhyming dictionary, where the rhyming words "Daenen", "gaehnen", "waehnen", and "Hyaenen" are located close together, and among similar words (except that accentuation is neglected). The dictionary was built automatically by sorting a medium set of german words with respect to reverse (i.e. right to left) lexicographic order of phonem equivalence classes. The latter had been ad-hoc determined as: {a} {aa,aah,ah,au} {ae,e,oe} {aeh,ee,eh,oeh} {aeu,ai,ay,ei,eu,ey,oy} {aeuh,aih,ayh,eih,euh,eyh,oyh} {auh} {b} {bb} {c} {cc} {ch} {ck,kk} {d} {dd} {dt,t,th} {f,ph} {ff} {g,gh} {gg} {h} {i,ue} {ie,ieh,ih,ueh} {j} {k} {l} {ll} {m} {mm} {n} {nn} {o} {oh} {p} {pp} {q} {qu} {r} {rh} {rr} {s} {sc} {sch} {ss} {tt} {u} {uh} {uy} {uyh} {v} {vv} {w} {ww} {x} {y} {ya} {yah} {ye} {yeh} {yh} {yo} {yoh} {yu} {yuh} {z} {zz}. Long (e.g. aeh, n) and short (e.g. ae, nn) phonem forms were distinguished; a short vowel before a long consonant was replaced by its long form before sorting (e.g. ae n → aeh n). For example, representing an equivalence class by its first member, "Daenen" and "gaehnen" was mapped to "n ae n aeh d" and "n ae n aeh g" for sorting, respectively.