Closed palemieux closed 4 years ago
The proposed image does not exactly match the note either. The note says:
the presentation processor selected half-width variants of the affected characters, but the resulting width is larger than the em square of the surrounding vertical text
I'm fine with using the proposed image but the note would have to be changed.
This seems to be a simple editorial change. I would propose to keep it in scope for TTML2 2nd Ed.
The intent of the right-most line of the sample rendition and the note was to depict the case where the implementation expands the line are in the block progression dimension to accommodate tate-chu-yoko text that extends outside the EM square (without reducing font size); therefore, I will correct the note to match the current sample rendition (rather than replacing the rendition).
@skynavga Having an example that does not follow the recommended behavior (half-width variant forms may be selected, a ligature may be selected, a smaller font size may be applied, etc. At a minimum, an implementation that supports this style property must be able to select half-width variant forms if available) is not a good idea. Why can't the rendering above, or one generated using TTPE be used?
I am missing something here?
The new text says:
In the above example, the rightmost region depicts the use of the all value, where the presentation processor has selected half-width variant glyphs for the affected characters, thus allowing the combined text to fit within same EM square as the surrounding vertical text.
The right-most image demonstrates more than selecting half-width variant glyphs. Some stretching is used also. I still find that confusing. I would either:
It is not the intent of the note to precisely prescribe the rendering behavior. In fact, being overly prescriptive is undesirable. The only thing the note needs to impart is that "something" having to do with "half width" is applied to "some" glyphs in order to make the text fit in the EM box. All else is a detail of the implementation. So, I would oppose any attempt to further tweak this language or example.
Propose this be re-closed without further action given the original commenter approved the already approved (and merged) PR (#1180).
The Timed Text Working Group just discussed Text Combine example is incorrect/misleading. ttml2#1128
, and agreed to the following:
SUMMARY: @skynavga to remove reference to "half width" for this example.
I addressed https://github.com/w3c/ttml2/issues/1128#issuecomment-551149035 in w3c/ttml2#1187 by retaining reference to half-width variants but indicating that it was a scaled form thereof. I hope this will suffice, as it is still desirable to refer to half-width variants given the prose
At a minimum, an implementation that supports this style property must be able to select half-width variant forms if available.
The note below the text combine example at https://www.w3.org/TR/ttml2/#style-attribute-textCombine implies that the render on the right uses different variants of the font for the affected characters. The latter however appear identical to the render on the left.
Suggest using the following render instead: