but for 90% of my writing I ain't bovvered
Jottings off the cuff - in between other writing :
If you do not want to chaotify your coded text to get your composed type correct, then you have to separate processing of character content data from glyph appearance data. The UTC and SC2/WG2 cannot encode glyph appearance data as character content data.
Either you should use a smart font in simple mode with the default shapings of the CMAP Character Map. In this case you get your default cursive lower case letter a for LATIN SMALL LETTER A, and you forget about whether you want this or that swash for setting the sonnets of Michelangelo.
Or you should use a smart font in smart mode and in this case you will discover that you can match the appearance of Early Modern German in glyph space and preserve your plain text plain in citations, and you can switch to Modern German and input your comment.
I went over the Unicode debates starting in 1991 and the TrueType debates starting in 1992. Nobody intended that you should have to bother with font fiddling the way you had to bother with font fiddling in Adobe applications - and still have to bother with font fiddling in buried menus.
Note that we essentially belong in the world of fraktur that knows only two letter cases and only one letter style. There is no small upper case, and there is no cursive. There is also no ligation as ligation is not set between syllables and compounds in German.
Svenska Akademien that awards the Nobel Prize for Literature began as an effort by Carl Linnæus, Lars Salvius and others to try to introduce spelling and setting more in agreement with the academies of England and France, and less with German printers in Sweden.
A century later Oxford University Press and Cambridge University Press essentially saw to it that the number of alphabet styles in manual composition was carried forward into mechanised composition, because they are useful in increasing the legibility of line layout.
If you do not separate processing of character content data from processing of glyph appearance data, then you cannot match the appearance of Harts Rules for Compositors and Readers at the University Press Oxford, the Oxford English Dictionary ...
German was different and German is still different. There are people, even people one would want to think of as intelligent, who post that they intend to insert control characters for ligation (ZERO WIDTH JOINER and ZERO WIDTH NON-JOINER) as well as insert combining diacritics.
As per ISO-IEC 10646:2003, if the above control characters for ligation are inserted, they affect searching and sorting. If the audience does not insert these control characters in the search string, the search is unsuccesful. Same with matching appearance in character space by combining diacritics - the plain text is not preserved as plain.
Nobody is saying that if you want to write a shopping list, you should use Linotype Zapfino. Everybody is saying that if you want to write a book with accepted Anglo-American academic typography, then you should use Apple Hoefler, Linotype Palatino, or Adobe Minion Pro (less the PUA that Adobe and Lino do not declare in their products).
And it's the same if you want to write a dictionary of Kiswahili with accepted Anglo-American academic typography as well as etymon in Farsi, Urdu, and Arabic. I could say the same of writing a dictionary of Portuguese or Spanish except that I wonder whether Arabic etymon are acceptable. I simply don't know.
hh
