On 7/7/2025 7:48 PM, Pablo Rodriguez via ntg-context wrote:
Language tagging should be able for some block elements (at least, delimitedtext and probably description).
That should not be too hard, depending on how specific one wants to be. It's more about choices than complexity of implementation.
The latest tagging sample mentioned in the list (shorter version, https://perspectivia.net/servlets/MCRFileNodeServlet/pnet_derivate_00007290/...) contains also some tagging that is really hard to believe it might be right, such as:
/P <>BDC 0.145 Tw 9 0 0 9 66.4399 67.6929 Tm [(e)13.6 (.)27.4 (g)16.2 (. Peter Frankopan)-7 (, )]TJ EMC /P <>BDC /TT1 1 Tf 10.022 0 Td [(The Earth Transformed: An )-0.6 (Untold History)]TJ EMC
At least, there are the following issues here:
1. Too many /P elements for what should be /Span elements. 2. “The Earth Transformed: An Untold Story” is English, not German. 3. /MCID are orphan (they lack their corresponding /StructElem).
#2 may be a human honest mistake, but #1 and #3?
My impression is that many (when not most) people may be struggling with proper tagging.
If it somehow validates then the objectives are probably reached. Which is what much tagging will boil down to in the end. One problem is that there are restrictions to what elements can nest which is why we need to play safe. In the average text we have just titles, itemization, simple tables, maybe simple math (most articles and text books actually are just that). But what if we start nesting, and do typographically interesting things. Say you have somethign that tags as label but it has an image in it. Or a (nested) link which has its own restrictions. If accessibility is really important one should just make a stripped own no-fancy-typesetting document alongside. After all, in that case looks don't matter. Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------