Primary identifiers for Internet Drafts

[strogonoff]

Currently, Relaton data provides document identifiers for Internet Drafts like this:

https://github.com/ietf-ribose/relaton-data-ids/blob/63bd977c0359878c956b9ad5d3850c33e55a88db/data/draft-ietf-intarea-tunnels-10.yaml#L14-L22

Note the pattern marked as “primary”. Presumably, this is the one suggested to be used when citing, and it has the format I-D ietf-intarea-tunnels.

From my understanding, IETF uses draft-ietf-intarea-tunnels when citing (Datatracker returns those, too). Do we want to include this identifier and indicate it as primary to conform to real-life use scenarios but violate Relaton’s wider schema? (Or am I misunderstanding something and I-D ietf-intarea-tunnels is in fact used?)

[ronaldtse]

@strogonoff draft-ietf-intarea-tunnels is definitely one "primary" identifier.

The pattern I-D.xxx is used in xml2rfc, which is the authoritative tool used to create IETF documents: https://www.ietf.org/proceedings/65/slides/xml2rfc-0.pdf

e.g.

<!ENTITY dos-draft PUBLIC "" 'http://xml.resource.org/public/rfc/bibxml3/reference.I-D.iab-dos.xml'> ]>

But I-D ietf-intarea-tunnels is probably not the primary identifier.

Ping @andrew2net .

[TonyLHansen]

I would call the -10 identifier as primary, with the value without the -10 as secondary. After all, the -10 identifier uniquely identifies a particular version of a draft.

The same would be true for the -09, -08, -07, etc versions.

But I'm not sure what connotations there are for the identifier to be primary.

[strogonoff]

The connotation, from what I understand, is just that authors can use that exact string to unambiguously reference/cite a document. (E.g., DOI and ISBN are also identifiers, but no one cites by DOI or ISBN, so Relaton data does not mark them as primary.)

[andrew2net]

But I-D ietf-intarea-tunnels is probably not the primary identifier.

@ronaldtse as I understand the primary ID is used to cite a document. The relatnon-* gems accept primary ID as an argument to get a document.

[ronaldtse]

I understand the primary ID is used to cite a document. The relatnon-* gems accept primary ID as an argument to get a document.

Relaton has to accept any ID (primary or not primary) used to cite a document. The "primary" tag is only useful when a reading application wants to display the ID. @andrew2net can you file a ticket at the appropriate place? Thanks.

[strogonoff]

I understand the primary ID is used to cite a document. The relatnon-* gems accept primary ID as an argument to get a document.

Relaton has to accept any ID (primary or not primary) used to cite a document. The "primary" tag is only useful when a reading application wants to display the ID. @andrew2net can you file a ticket at the appropriate place? Thanks.

I thought Relaton enforces primary usage. If it doesn’t, it would allow authors to cite by DOI or ISBN, and from my understanding this should not be the case.

[ronaldtse]

I thought Relaton enforces primary usage. If it doesn’t, it would allow authors to cite by DOI or ISBN, and from my understanding this should not be the case.

This is a misunderstanding in terminology, I should have been more clear.

• Relaton should allow location/search of a bibliographic item by any identifier (including DOI)
• The user application (this is the application that renders a bibliographic reference) should show the primary identifier as the main identifier.

[strogonoff]

I agree on search, I meant using any non-primary docid when citing/referencing.

[andrew2net]

• Relaton should allow location/search of a bibliographic item by any identifier (including DOI)

@ronaldtse most data sources allow search only by PubID. I'm working on the relaton-doi now, which will allow fetching documents by DOI from the Crossref service. What other identifiers do we need to support?

[ronaldtse]

@andrew2net For data sources that we need to "search on an external service" it would be difficult to allow searching using other criteria. However, for the data sources that we have a full dataset, we can create indexes using those other kinds of IDs too.

[ronaldtse]

@strogonoff is this task still valid? The anchor used by I-D is correct as per here: https://github.com/ietf-ribose/bibxml-service/issues/151#issuecomment-1064218979

[strogonoff]

@ronaldtse I think this issue has been superseded by https://github.com/ietf-ribose/relaton-data-ids/issues/14.