# Template Key Justification
The following describes how i18n keys are justified in a html source file:

+ Traverse the parsed html tree:
    + if a node is whitelisted,
        + raise diagnostic if node contains mixed content (tags and non-whitespace text).
        + try parsing the original i18n attribute and add to the list of known keys.
    + else
        + raise diagnostic if node contains non-whitespace text
        + raise diagnostic if node has an i18n attribute.
+ Start a new postfix counter at zero.
+ Function for generating a unique i18n key (with an optional original key):
    + if no key is provided or a similar key has been generated by this function,
        + generate new keys until it is not in the list of known keys.
    + add the key to the list of known keys.
    + add the key to the list of keys generated by this function.
    + return the key.
+ For all whitelisted nodes from the tree traversal:
    + create a new i18n attribute.
    + if content localization is allowed,
        + get the preferred key from the original attribute.
        + if the node has text content or the preferred key exists,
            + generate and store a unique key in the new i18n attribute.
    + else, if the original attribute has a key for html or text content
        + raise diagnostic that already localized content is not allowed
    + else, if the node has text content
        + raise diagnostic that content is not allowed
    + for each whitelisted attribute name that exists:
        + get the preferred key from the original attribute.
        + generate and store a unique key in the new attribute.
    + for each attribute name (non text/html) in the original key:
        + raise diagnostic if the name is not whitelisted.
    + if the new i18n attribute is empty, delete the original attribute
    + else, set replace the original attribute.

## Additional notes
+ Html attributes named `"html"` or `"text"` are ignored due to the way aurelia-i18n works.