Thu 20 Jul 2017 03:31:10 PM UTC, comment #4:
I will describe the ideal, then describe how to approximate that in a monospace font.
There are two ways to write the silent iota in Greek. The iota adscript is a normal iota which the reader just has to know is silent. The iota subscript is a hook below the preceding vowel.
Unicode provides both forms. The iota subscript is U+0345 COMBINING GREEK YPOGEGRAMMENI. The iota adscript is U+1FBE GREEK PROSGEGRAMMENI. The iota adscript is also U+03B9 GREEK SMALL LETTER IOTA: they are canonically equivalent, because they look, behave, and are identical.
One might say that there must be a difference or they wouldn’t both have been encoded. The answer is that ELOT requested a bunch of unnecessary characters which Unicode accepted to effect the merge with ISO 10646. Encoding both was like encoding the Latin ⟨h⟩ twice because French has both h aspiré and h muet.
FreeMono’s glyph for U+1FBE is a spacing low hook. This is, as far as I know, unattested.
So what should it look like? A small iota. It should look and behave 100% identical to U+03B9. One might say that that would obscure the difference between U+0345 and U+1FBE. My response is that the only difference is an artificial artifact of the encoding, not a real difference in the Greek script.
> I mean, one has the options of squishing the glyphs to fit the bounding box, or having them spill out badly.
That is not true. No squishing or spilling is required. In fact, no visual interaction between U+1FBE and the preceding vowel is required. U+1FBE is not a combining character: it stands on its own; in a monospace font, it goes in its own cell.
I’m not talking about characters like U+1FBC GREEK CAPITAL LETTER ALPHA WITH PROSGEGRAMMENI, where I agree that the iota adscript would require some unpleasant squishing: I am only referring to the stand-alone U+1FBE.
|