How to encode emphasis markers interspersed with gloss

[knagasaki]

from: martindholmes/rubyForTEI#10 (comment)

I tried to make an example of a markup using <note> with emphasis markers under <sourceDoc>. Given that encoders can not determine the meaning of the markers, I suggest to use <hi>, if it will be independently encoded. As the targeted text of <note> is indicated as pseudo-standoff, the <hi> for the emphasis markers can be embedded easily. The URIs of the image are inserted compliant with IIIF. It is just an attempt. I hope your comment about not only usage of tags but also entire strategy of encoding.

https://ids.lib.harvard.edu/ids/iiif/26478854/full/1800,/0/default.jpg

   <sourceDoc source="https://iiif.lib.harvard.edu/manifests/drs:26478660">
      <surface
         source="https://iiif.lib.harvard.edu/manifests/drs:26478660/canvas/canvas-26478854.json"
         facs="https://ids.lib.harvard.edu/ids/iiif/26478854/full/1800,/0/default.jpg">
         <line>
            <hi style="white-circle-right">而較之可手遂<note place="right"
               targetEnd="#note0001e">原評󠄁令有以自容</note>命𥁞絶其纓</hi>
            <hi style="sesame-right">尋</hi><anchor type="noteEnd" xml:id="note0001e"/>荘王與𣈆<hi
               style="sesame-right">戰</hi>一人 </line>
      </surface>
   </sourceDoc>

[duncdrum]

here was my encoding from the other repo

<pb n='三' facs="Images/26478854.jpg"/>
<line style='writing-mode: vertical-rl'>
  <ab rend='○' style='text-emphasis-style: open'>
    而較之可乎遂
     <r:ruby>
       <r:rb rend='○' style='text-emphasis-style: open'>命書絶其纓</r:rb>
       <r:rt>原<g>⿰言仐</g>令有以自</r:rt>
    </r:ruby>
  </ab>
          <!-- I m not to happy about having to split here, but i don't see another option. 
    this is where overlap problem will occur if i want to mark the semantic unit of the text across 自容 or 纓尋 boundaries
    -->
  <ab>
     <r:ruby>
       <r:rb>尋</r:rb>
       <r:rt>容</r:rt>
     </r:ruby>
    荘王與𣈆
  </ab>
</line>

still seems more concise, ideally we could avoid repeating @rend and @style on rb but for that we would need inheritance rules, that include rb but exclude rt

[knagasaki]

I think style='text-emphasis-style: open' in the first <ab> should be inherited in the <r:rb>. In that case, if a series of emphasis-marks is stopped in the middle of <r:rb>, the encoding will be more complicated. Then, I recommend to use <hi>, <emph>, or something like that with @rend, while annotated elements like <note> or <r:ruby> are marked up in the manner of standoff. And then, I would like to ask you to use another example for this. The example includes two difficulties. I would like to discuss the two in another place. One is usage of <g>⿰言仐</g>. Another is to add @type to <r:ruby>.

[martindholmes]

@knagasaki I wouldn't use @rend, I would just use another @style (or @rendition pointing to a <rendition> element.

[knagasaki]

@knagasaki I wouldn't use @rend, I would just use another @style (or @rendition pointing to a <rendition> element.

@martindholmes , I might not understand enough the usage of @rend. Is it used only for simple (or pre-defined) style such as underline or something like that? Anyway, I don't oppose to use @style there. My point is to use an attribute for it.

[martindholmes]

@knagasaki @rend is an old attribute whose values are not controlled in any way, so every project has to define their own usage, and encoding is not easily interoperable. @style uses the well-defined, well-established CSS standard, so everyone knows what @style values mean, and they can easily be handled when creating HTML from your TEI. @rendition is a pointer to a <rendition> element in the header, whose content is almost always CSS, so it's a way of defining a style ruleset once, and then using it many times.

So I think I would only use @rend if the style I'm encoding doesn't have any representation in CSS at all. That doesn't often happen these days. :-)

[knagasaki]

@knagasaki @rend is an old attribute whose values are not controlled in any way, so every project has to define their own usage, and encoding is not easily interoperable. @style uses the well-defined, well-established CSS standard, so everyone knows what @style values mean, and they can easily be handled when creating HTML from your TEI. @rendition is a pointer to a <rendition> element in the header, whose content is almost always CSS, so it's a way of defining a style ruleset once, and then using it many times.

So I think I would only use @rend if the style I'm encoding doesn't have any representation in CSS at all. That doesn't often happen these days. :-)

@martindholmes, thank you for the kind explanation. I understand the situation. I will use @style and @rendition from now.

[martindholmes]

The question of rt inheriting style from its ancestors can be overcome by using a global rendition element in the header:

<rendition selector="rt"> [...all the style for rt elements goes here...]</rendition>

However, I don't believe the Guidelines express clearly enough yet what the inheritance expectation is in these cases; does a rendition with @selector override styles declared directly on ancestor elements in the text? This could presumably be clarified by using !important, of course.

[duncdrum]

The problem with inheritance is that , correct me if I m wrong, the ruby element inherits the @Style from ab This is wrong rt should be excluded.

Repeating @Style on rb Is intended to help deal with the overlap. But fails to do so since rt still inherits from ab A rule that clarifies style inheritance for ruby would make for leaner markup, by allowing me to skip repeating a declaration on rb. If rt Were to not inherit style by default it would also not inherit the writing mode from line, so not ideal either.

So we need a solution that’s unambiguous as to which characters have emphasis markers (and which don’t) hopefully avoiding unnecessary repetition of attributes.

Something along the lines: Style inheritance stops at ruby Including rb, rt only inherits document wide properties / styles from non direct ancestors of ruby

[martindholmes]

Since the CSS text-emphasis spec is now pretty solid, and allows you to specify the character used, @MegJBrown, @martinascholger, @hcayless and I think that the next stage would be to encode two good examples using CSS text-emphasis, and run them by everyone before incorporating them into the Guidelines.