NeXML generated by OntoTrace uses HTTP URIs as document-local identifiers

[hlapp]

The NeXML generated by OntoTrace includes elements such as the following:

<otu id="http://purl.obolibrary.org/obo/VTO_0036225" label="Ictalurus punctatus" about="#http://purl.obolibrary.org/obo/VTO_0036225">

Note the hash (#) preceding the HTTP URI in the about attribute. Is this intentional, and why is it needed?

[balhoff]

The OTU has the URI as an ID as a convenient document-unique ID which also references to the taxon. The about is there to say what the <meta/> inside refer to, which with the hash is the element with that ID in the current document. Is it causing a problem somehow?

[balhoff]

I downloaded Oxygen (which I have alway found to be the most reliable XML tool) to check these. It does look like the URIs are invalid ID values, specifically due to the colons and slashes. These IDs are only used in OntoTrace and not in Phenex. This does need to be fixed.

[rvosa]

The hash is intentional and needed, conform the first bullet item example in this document: http://www.w3.org/2006/07/SWD/RDFa/syntax/#id103742

Basically, it is needed to reconcile the identification of the subject in the RDFa triple with the equivalent XML element.

_Edited to add:_ The hash is needed for relative URIs that are expanded on the xml:base. With absolute URIs that aren't in the same document scope this is not needed - so in principle the about attribute can do without - but then you still run into the problem of the invalid id attribute, which CANNOT be a URI. So, to ensure that both the XML is valid, and the RDFa subject is reconciled with the equivalent XML element, the element must have an NCName ID, and the about attribute must reference it, using the hash to indicate that it's a reference to something within the same document scope / base URI.

[balhoff]

Fixed by 201e8da8c214b80e5f906f5cdb01d5dac1552c09.