When I see Unicode-hacks like this (is it a hack? is Unicode supposed to do this?), my instinct reaction is that maybe Unicode is bloated beyond a reasonable mission scope. Maybe it tried to do too much, or maybe it just made some mistakes along the way.
But I don't really know much about it, and I'm suspicious of said instinct. HN what are your thoughts -- what are some controversial ideas around Unicode, and would there have been any simpler alternative that could make sense?
Some languages must naturally be composed of predictable, combinable marks to form glyphs. There are many separated diacritic codepoints, like "\u0061\u0301" = à.
I don't think this "hack" points to Unicode doing too much.
For questionable technical design decisions, look at UTF16, surrogate pairs, etc.
The 𝐛𝐨𝐥𝐝, 𝑖𝑡𝑎𝑙𝑖𝑐, 𝕕𝕠𝕦𝕓𝕝𝕖 𝕤𝕥𝕣𝕦𝕔𝕜, and 𝔟𝔩𝔞𝔠𝔨𝔩𝔢𝔱𝔱𝔢𝔯 gylphs are in Unicode for mathematics, where the style of glyphs is not a merely stylistic choice but is part of their meaning.
e.g.
ℝ denotes the set of real numbers, whilst R denotes the gas constant. 𝐑 would by convention be the name of some matrix.
ln is the natural logarithm, 𝑙𝑛 is the product of 𝑙 and 𝑛. (Only a monster would actually write ln 𝑙𝑛, though.)
I don't know why the consortium chose to include these glyphs rather than push it out to markup languages, though.
There is always what we had before: ASCII + Codepages, 128 standard fixed chars and 128 "extended" chars dependent on the selected Codepage. We could've just agreed on a control character to use as an escape for switching codepage and called it done. That wouldn't have given us such robust support for diacritics however.
26 comments
[ 3.9 ms ] story [ 64.9 ms ] threadon a side note, attempted triforce is attempted
̲̞̫̾̈́͑́H̜̆͢Ě͕͔̫̹̖͇̇͊̚ ̨̮̍̆̋̑ͧČͥ͌̽̾̚Ō̪̠̟̙̟͌͑̊͡M͙͇̙͎͉̏ͨ͊͒̂ͨͅẸ̷̩͕̞͚ͣͤ̉ͣ̓S̠͖̣̞̽ͩ́ͦ̔̏̚!͖͇̯̙̙̘̥ͥ̎̋͊
http://www.sprezzkeyboard.com/
But I don't really know much about it, and I'm suspicious of said instinct. HN what are your thoughts -- what are some controversial ideas around Unicode, and would there have been any simpler alternative that could make sense?
I don't think this "hack" points to Unicode doing too much.
For questionable technical design decisions, look at UTF16, surrogate pairs, etc.
e.g.
ℝ denotes the set of real numbers, whilst R denotes the gas constant. 𝐑 would by convention be the name of some matrix.
ln is the natural logarithm, 𝑙𝑛 is the product of 𝑙 and 𝑛. (Only a monster would actually write ln 𝑙𝑛, though.)
I don't know why the consortium chose to include these glyphs rather than push it out to markup languages, though.
There is always what we had before: ASCII + Codepages, 128 standard fixed chars and 128 "extended" chars dependent on the selected Codepage. We could've just agreed on a control character to use as an escape for switching codepage and called it done. That wouldn't have given us such robust support for diacritics however.