Their function is to convey what kind of hOCR/HTML markup is to be expected, so it's kind of like a set of capabilities with additional description of the purpose or origin of the data.
If profiles are conceptually distinct from the HTML markup restrictions (formats?), we could introduce a ocr-profile metadata field or similar and create a base list of profiles, including samples. If profiles and formats are related enough, they should be merged (#62)
Their function is to convey what kind of hOCR/HTML markup is to be expected, so it's kind of like a set of capabilities with additional description of the purpose or origin of the data.
If profiles are conceptually distinct from the HTML markup restrictions (formats?), we could introduce a
ocr-profilemetadata field or similar and create a base list of profiles, including samples. If profiles and formats are related enough, they should be merged (#62)See also https://github.com/kba/hocr-spec/issues/17#issuecomment-256117486