fred-wang
commented
4 years ago consensus from 2020/06/23: postpone to a future version
Open fred-wang opened 5 years ago
fred-wang
commented
4 years ago consensus from 2020/06/23: postpone to a future version
Just importing this from gitlab, but I think it's not a priority for now.
"I'm opening this for a future version of the implementation note. The fonts made by the GUST e-foundry group do not contains any kerning values for scripts. The MathKernInfo table provides additional kerning corrections when positioning scripts. The choice of kern values depend on the vertical offsets of scripts and so it seems that it will make preferred widths depend on vertical metrics of MathML boxes. However, these kerning rules only apply in the simplified case where the base and scripts are just single glyph (not complex subformulas) and so it's possible to determine the vertical shifts immediately during the "preferred width calculation" phase. Some cases to consider:
Base with many pre/post scripts (mmultiscripts) Base/Scripts that contains a single glyph nested in mrow-like elements"
Original report: https://gitlab.com/mathml/MathMLinHTML5/issues/11