w3c
AccName should account for images that include alternative text in the file contents
AccName should account for images that include alternative text in the file contents
For example, if an HTML alt attribute is not provided, but the image file includes an embedded EXIF description, that could be used. Likewise, SVG <title> could be used in place of @alt regardless of whether there are accessible grouped contents in the SVG.
I'm envisioning a bullet point in the alt text computation that points out to another resource that contains file format-specific info. SVG -> title and/or descendant contents; HEIF -> altt; GIF -> …
HEIF (HEIF, HEIC, AVIF, etc.) defines its form of embedded "Accessibility Text" in section 1.1.1. I'm not sure what the best public link is for HEIF, but here's one public resource that mentions altt https://nokiatech.github.io/heif/technical.html
When embedded in html I think we need to start with html-aam. We could then clarify 2D to include the ability to get the text alternative from within a source and not only from Elements and attributes.
When embedded in html I think we need to start with html-aam.
Would you still consider it appropriate to define the behavior in HTML-AAM when an author could use SVG to embed a raster/bitmap <image> in a context outside HTML? Admittedly this example is slightly contrived, but worth considering.
<image width="100" height="100" xlink:href="./bitmap.png"></image>
IPTC Photo Metadata Standard 2021.1 now includes http://www.iptc.org/std/photometadata/specification/IPTC-PhotoMetadata#alt-text-accessibility and http://www.iptc.org/std/photometadata/specification/IPTC-PhotoMetadata#extended-description-accessibility