[bact]

Add paragraphs to say

• We publish spec text and machine-readable artifacts (ontology/SHACL/schema). They intended to agree.
• If normative text and artifacts disagree, it is a defect; until it resolves, the text wins.
• Some requirements in text are not yet implemented in the artifacts, so passing validation does not automatically assumed conformance.

[goneall]

Although I generally agree with this approach - especially making it clear to the reader what should take precedence - there have been several cases in the past where the OWL document was correct and we updated the text.

I think it would be better to state something like - "any conflicts between the artifacts and text should be treated as a defect to be resolved. Until resolution, the text will take precedence." - or something similar.

[bact]

I think it would be better to state something like - "any conflicts between the artifacts and text should be treated as a defect to be resolved. Until resolution, the text will take precedence." - or something similar.

Revised as suggested.

[bact]

@henrylyons Please help review. Thank you.

[zvr]

While it's useful to tell people what should prevail in the case of conflicting information, I firmly believe that this should not be placed inside the official specification text.

The specification that we publish is the text; the fact that we also produce other artifacts (ontology, schema, or even code libraries that read and write data) does not imply that these should be regarded as a higher source of truth. If there is need, as Gary writes, to update the text, we do so.

Additionally, I am reluctant to treat conflicts as "defects" that should be resolved. We know there are a number of things expressed in the text that will never be translated into the other forms. Typical example, for more than a dozen years, is the valid strings for SPDX License Expressions. Or, a more recent one, country names.