Dear @michaelnmmeyer , this is not urgent, plenty enough to address it sometime after you return from leave, but I am noting this now so it doesn't get lost.
Now that the batch adjustment of the existing files has been completed, I think it is time to restrict the permitted values of @type on <g> to what we actually use, namely (copy-pasted from the glyph management gdoc, which everyone has at least tacitly accepted):
- “symbol” will be the default value, for generic symbols, meaning that no assertion is made regarding function
- “punctuation” for graphemes (functional marks) expressly interpreted as punctuation marks in the strict sense
- “connector” for graphemes (functional marks) expressly interpreted as connectors
- “ideogram” for graphemes expressly interpreted as ideographic
- “numeral” for graphemes expressly interpreted as numeric, encoded as before and not discussed here
- “phonographic” as a provision for the future, not for use in our files, but in case someone at some point wants to use for phonographic glyphs
It seems that some or all values of @type still look up the authority file and generate display using the mapping there, e.g. <g ref="sym:bullet-high" type="pc">,</g> gets displayed as
I would strongly prefer to permit only the above six values (thus allowing schema-based suggested values and reducing human error). Thus, encoding such as <g ref="sym:bullet-high" type="pc">,</g> should be flagged as an error.
If that is not feasible, and the encoding with @type containing a token is absolutely essential for TST or some legacy purpose (which @manufrancis may be able to clarify), then the display should be changed so that when @type and @ref are both present, it is the @ref that determines the display overriding the @type, and not the other way around.
Dear @michaelnmmeyer , this is not urgent, plenty enough to address it sometime after you return from leave, but I am noting this now so it doesn't get lost.
Now that the batch adjustment of the existing files has been completed, I think it is time to restrict the permitted values of
@typeon<g>to what we actually use, namely (copy-pasted from the glyph management gdoc, which everyone has at least tacitly accepted):It seems that some or all values of
@typestill look up the authority file and generate display using the mapping there, e.g.<g ref="sym:bullet-high" type="pc">,</g>gets displayed asI would strongly prefer to permit only the above six values (thus allowing schema-based suggested values and reducing human error). Thus, encoding such as
<g ref="sym:bullet-high" type="pc">,</g>should be flagged as an error.If that is not feasible, and the encoding with
@typecontaining a token is absolutely essential for TST or some legacy purpose (which @manufrancis may be able to clarify), then the display should be changed so that when@typeand@refare both present, it is the@refthat determines the display overriding the@type, and not the other way around.