Re: Policy on character name aliases?

From: Asmus Freytag (
Date: Tue Mar 09 2010 - 22:32:18 CST

  • Next message: Andrew West: "Re: property, character, and sequence name loose matching"

    On 3/9/2010 3:59 PM, karl williamson wrote:
    > It appears that there is a high bar to getting an alias made for a
    > character name. Why is that? It's very clear that the original name
    > must be immutable, but why not add an alias when it's shown that the
    > original was an unfortunate choice?
    "Unfortunate" is an unfortunate term here, because it's rather open ended.

    Aliases are costly. Every time you add one, many, if not all users of
    the standard need to be aware of it, because the names (and aliases) are
    supposed to act not as mere descriptions, but as formal identifiers for
    the character.

    A better standard for adding an alias would be the formal recognition
    that the current name "is defective as a formal identifier". A
    misspelled word in a character name, by precedent, counts as such a defect.

    A mere preference or a difference in interpretation does not. In the
    early days of the standard the word ligature was replaced by the word
    letter for a few character names because of such preference issues
    raised by a small number of NBs. To prevent such issues, names were made
    immutable. As a result, a lot of work was able to be carried out without
    distractions based on minor naming preferences. Adding a backdoor to
    that with (formal!) aliases would be counterproductive.

    Some character names can't be fixed with aliases. Think of the tacks
    where up and down were applied inconsistently. Adding aliases with the
    opposite directionality would only confuse matters - the current
    solution, i.e. to document the issue, creates greater clarity.

    Sometimes there's a blatant misidentification of a character. These
    cases are rare. Think of the OY which really is a GHA. In those cases,
    an alias can be helpful. However, they would satisfy the criterion of
    the name being a defective identifier, because the name is unrelated to
    the character at hand (as opposed to merely not perfectly related to it,
    in someone's view).


    This archive was generated by hypermail 2.1.5 : Tue Mar 09 2010 - 22:37:16 CST