Re: Metadata question

Hey John,

I looked at the EPUB 3.3 spec[1] because (we hope) that spec is more readable than EPUB 3.2 (though technically backward compatible) and, maybe even more importantly, if there are things that require improvement, that is the place to do it...

The definition of identifier-type[2] says that it may "extend" (ie, used in a "refine" structure) two properties: dc:identifier and dc:source. This means that epubcheck correctly flags your code. In other words, the error is not related to the fact of cross-linking schemes but, rather, that identifier-type cannot be used on dc:subject. Mixing schemes is, by itself, not an issue: there is an example in [3] showing the usage of MARC terms with respect to a refine.

Of course, if you feel that there is a missing functionality for metadata usage, then this should be discussed within the framework of the EPUB 3.3 development. The EPUB 3.3. issue list[4] is your friend!

I hope this helps,

Cheers

Ivan


[1] https://w3c.github.io/epub-specs/epub33/core/ <https://w3c.github.io/epub-specs/epub33/core/>
[2] https://w3c.github.io/epub-specs/epub33/core/#sec-property-identifier-type <https://w3c.github.io/epub-specs/epub33/core/#sec-property-identifier-type>
[3] https://w3c.github.io/epub-specs/epub33/core/#sec-property-role <https://w3c.github.io/epub-specs/epub33/core/#example-78>
[4] https://github.com/w3c/epub-specs/issues/ <https://github.com/w3c/epub-specs/issues/>


> On 23 Jun 2021, at 23:48, John Foliot <john@foliot.ca> wrote:
> 
> Hello All,
> 
> As a long time accessibility specialist who's been involved in W3C work for more than 15 years, I am now getting involved with accessibility in ePub (thanks to a new contradict), and while accessibility principles are old-hat to me, some of the peculiarities of epub packaging are still kind of new to me.
> 
> While I hope to get more involved overall with epub+accessibility, at this time I wonder if I may pose a question to the assembled experts on this list? Specifically:
> 
> Currently epubcheck is telling me that there is an error with some of the metadata in a specific opf file, but epubcheck doesn't tell me with enough detail what the actual error is, only where to find it (sigh). This is the problematic error message and metadata:
> 
> "Error while parsing file: Property "identifier-type" must refine an "identifier" or "source" property"
> 
> <dc:subject xmlns:html="http://www.w3.org/1999/xhtml <http://www.w3.org/1999/xhtml>" id="subject01">Marketing</dc:subject>
> <meta xmlns:html="http://www.w3.org/1999/xhtml <http://www.w3.org/1999/xhtml>" refines="#subject01" property="identifier-type" scheme="onix:codelist26">10</meta>
> 
> My *assumption* at this time is that the issue here is that this code is attempting to refine the dc:subject with an onix 'term', but... (?) 
> 
> Questions: 
> Can Dublin Core and Onix metadata live together in harmony in the opf file?
> Am I correct in my assumption that the actual error is the "cross-linking" of the two schemes?
> ...and would the remediation path be to remove the refines="#subject01" entry from the second declaration?
> Or am I completely missing the boat and looking in the wrong place? Is this correct in principle, but incorrect in implementation?
> Any light anyone can shed on this would be appreciated - references would be an absolute bonus - but today I'd be happy to just understand what to do to fix this.
> 
> Thanks in advance!
> 
> JF
> -- 
> John Foliot | Senior Industry Specialist, Digital Accessibility
> 
> "I made this so long because I did not have time to make it shorter." - Pascal
> "links go places, buttons do things"


----
Ivan Herman, W3C 
Home: http://www.w3.org/People/Ivan/
mobile: +33 6 52 46 00 43
ORCID ID: https://orcid.org/0000-0003-0782-2704