For tags, switch from Unicode “letter” to “alphabetic” for allowed characters?

[jotaen]

A “letter” (the allowed character class for tags) is currently defined as “character as defined by the Unicode Letter category (L)”.

For [x]it! (a file format for check lists), there was a discussion on HackerNews in which one commentator suggested that it would be better to use the Unicode “alphabetic” category instead, which appears to be a superset of “letter”.

This topic applies to klog in the same fashion, so depending of the outcome of that investigation in [x]it!, the definition of “letter” might need changing here as well.

Since “alphabetic” would be a superset of “letter”, this change would be backwards-compatible.

[jotaen]

tldr; I’m closing this for now, as I think the current definition is sufficient, especially considering that the “letter” definition is only relevant for tags, but not for any other text.

The Unicode categories are defined as:

• Letter: Lu + Ll + Lt + Lm + Lo
• Alphabetic: Lu + Ll + Lt + Lm + Lo + Other_Lowercase + Other_Uppercase + Nl + Other_Alphabetic

So, “alphabetic” is a superset of “letter”, that adds Other_Lowercase + Other_Uppercase + Nl + Other_Alphabetic:

• Other_Lowercase and Other_Uppercase contains a variety of miscellaneous characters such as roman numerals or letters/numbers with circles around it.
• Letter_Number contains “letterlike numeric characters”, mostly ancient ones, like roman, old persian, cuneiform, etc.
• Other_Alphabetic mostly contains code points from the “non-spacing mark” and “spacing combining marks” categories.

For tags, I’m not sure that any of these additional categories would be useful.