On 5/6/23, G. Branden Robinson <g.branden.robin...@gmail.com> wrote: > I disagree with this too. Part of the value of encapsulation of the > fundamental character type inside a formatter-specific type is that we > can change our minds _again_ if circumstances warrant.
Good point, and that moots all my points in this email. Nonetheless, I'll clarify one of them: > At 2023-04-29T22:33:52-0500, Dave Kemper wrote: >> and a theoretical future chapter-at-once algorithm (to, for example, >> optimize page layouts to eliminate widows) > > Well, if you format each paragraph in a diversion, you don't need to > expand the formatter's view of the present as much. I was thinking of (but didn't say out loud) an algorithm, as far as I know nonexistent at present, that would prevent widows while keeping the bottom margin on every page the same. In practice, this can entail tweaking a paragraph on page 4 to prevent a widow on page 6. An algorithm that can't see past a single paragraph cannot accomplish this. Possibly no algorithm can accomplish this; it's a much harder problem than the one Knuth-Plass tackles, and might be getting into AI territory, which will automatically incur Doug's wrath.