Copying Greek Text Bug (?)

[darrylmy]

Now this is probably really nit-picky, and I'm not sure exactly if this is a bug, but it seems to be related to how Accordance (13.3.2) is copying text and rendering it in unicode.

 

After copying text from Accordance in unicode, the acute accent (apart from a breathing mark) is at the wrong angle when the font is changed to Minion Pro. The acute accent is almost vertical. (For reference, I am copying using unicode and the SBL Greek font from Accordance before changing the font to Minion pro.) When typing Greek manually, the accent is far less vertical. Here is an overlay of an example of the two different angles of the accent (yeah, I know, like I said, "nit-picky," but it's noticeable when I have portions of copied Greek text and typed Greek text in the same document):

 

 

I was guessing this is a copying/unicode issue because when I copy text from the NA28 from https://www.academic-bible.com/en/online-bibles/novum-testamentum-graece-na-28/read-the-bible-text/ and change the font to Minion Pro, the acute accent is at the correct angle.

 

I know this is a minor issue and I'm probably the only one affected by it, but could this "bug" be looked into?

 

Thanks,

 

Darryl

[jlm]

This is probably not a bug, but another undesirable result of Accordance converting Greek text to the standard decomposed Unicode form before placing it on the clipboard. In the past, the Unicode consortium thought that the tonos of monotonic orthography should be distinguished from the oxia of polytonic orthography, and not a few font designers thought they should be visually distinguished by angle. Now the Unicode consortium has realized that they are the same, and so the standard decomposition procedure turns oxia into tonos. Because Minion Pro was designed to distinguish these accents, you see the angle change. Based on today’s knowledge of Greek accents, the font is at fault, because it shouldn’t distinguish these accents visually. However, if Accordance wasn’t doing that standard decomposition, it wouldn’t be provoking an error on the part of the font. The decomposition is technically correct, but there’s a lot of software out there that’s not prepared to deal with it.

 

Reference: https://en.wikipedia.org/wiki/Greek_diacritics#Unicode

[darrylmy]

Thank you @jlm for explaining this to me. I went down the rabbit hole on this one to get my head around it and from what I can gather, the unicode standard is that oxia and tonos should be indistinguishable as you mention. However, when the character is decomposed the result should be the slanted acute accent mark (i.e., practically appear as an oxia although it is decomposed as a tonos). So if I'm understanding you right: when the character is decomposed by Accordance it is decomposing to tonos, which should be displayed as a slanted acute mark (i.e., appear as an oxia), but since Minion Pro distinguishes tonos, font doesn't get decomposed correctly. (It seems this can also vary among different websites with accented Greek text as well.)

 

So it seems that my options (for using Minion Pro) are: (1) wait and hope that Accordance is updated to handle fonts with distinguishing unicode ranges for oxia and tonos; or (2) modify Minion Pro by removing the tonos characters which have equivalent oxia characters.

[jlm]

I wrote the above without actually testing what Accordance was doing. It turns out there's another step: both tonos and oxia are considered equivalent to an acute accent, so what you're actually getting in the standard decomposed form (NFD) is a combining acute accent, character U+0301, and that's what's appearing at a different angle from the Greek accent you're typing.

 

If you're working in a plain text format like Markdown or LaTeX, all you have to do is apply a consistent Unicode normalization (NFD or NFC) to the whole file, and all your accents will come out the same. I normalize my files with a Python script.

 

If you're using a word processor, and it has kept the decomposed form as it was pasted, then you might be able to search your document for the combining acute accent (hexadecimal code 0301) and replace it with a combining oxia (hexadecimal code 1FFD). On macOS, the Character Viewer (a.k.a. Show Emoji & Symbols) will let you pick these characters if the search and replace dialog doesn't accept hexadecimal codes.