> It's my serious opinion in this case that a user > wanting to create documents with a nice quality mix > of curved and straight quotes that they should have > their fonts set up to handle it.
It is worth keeping in mind that the Linux problem with typographic quotes is not that of fonts, it is one of encoding. AFAIK, Latin1 does not have single typographic quotes codepoints (bad for the UK user!) and Latin2 does not have either single or double; no one is going to be running en_GB.cp1252 just because AbiWord messes up with quotes. My point in the original example was that she may not want to use curved quotes, but that the double remapping obscures from her that she is doing so; I am not saying that remapping to something more meaningfull than the undefined glyph symbol is not a good idea, merely that it has to be done in a way that the user can always tell that what she is seeing has been remapped from a non-existent glyph, that What You See Is Not What You Have. It seems to me that a best way to do this would be to have a separate fp_AbsentGlyph class of run which would contain a static remapping table for common glyphs, such as the typographics quotes, and in screen context it would draw in a distinct manner, say yellow on dark blue background (it would print normally, naturally). > In fact now that you've got me thinking about it, it > makes a whole lot of sense to put up a warning dialog > when a user tries to enable smart-quotes on a system > that can't physically display them. This might be a good idea, but it does not address the problem I am really concerned about, which is that by destructive, non-optional, translation of the quote glyphs you are making the document non- portable. > Also if you're doing Greek and > this is the right way to enter a breathing mark, the > language-specific smart quotes code should know about > it. I was not thinking of her typing in the Greek alphabet, I was thinking of her transliterating Greek using Latin1 characters, which is fairly standard in academia. Our fault here is not that our smart quote algorithm will get the translation wrong, she is using the quote in a way that makes it impossible for us to make the correct choice; our fault would be that the double remapping would hide from her the fact that we made error. I see no simple way around this scenario, even if the remapped character is displayed differently on screen, it would merely indicate that the contents of the document are different from the screen, not that they are wrong. Tomas
