bibxml-doi (bibxml7): target URL should use https and doi.org

[ajeanmahoney]

Describe the issue

The International DOI Foundation prefers that "https" and "doi.org" be used for DOI URLs. Please see the following DOI Factsheet: https://www.doi.org/factsheets/DOIProxy.html

Currently bib entries in bibxml-doi (bibxml7) use "http" and "dx.doi.org" in the URL of the target attribute of both the <reference> and <format> elements. For example:

https://bib.ietf.org/public/rfc/bibxml7/reference.DOI.10.1007/3-540-60865-6_43.xml Current: target="http://dx.doi.org/10.1007/3-540-60865-6_43" Preferred: target="https://doi.org/10.1007/3-540-60865-6_43"

Code of Conduct

• [X] I agree to follow the IETF's Code of Conduct

[ajeanmahoney]

Hi all, is there a status update on this? Thanks!

[rjsparks]

@ronaldtse is this something that can be changed?

[cabo]

kramdown-rfc is not currently providing target attributes for DOI. Is this something that should be changed? (The target should be a link from the DOI.xxx text instead of duplicating the text, which I'm not sure can be expressed.)

[ronaldtse]

@ajeanmahoney @rjsparks @cabo apologies for not tracking this sooner. We can certainly update the URLs.

Ping @strogonoff @stefanomunarini @andrew2net to update target URL.

[cabo]

The correct solution is to remove the target from the bibxml, as xml2rfc then inserts a correct target.

[rjsparks]

The correct solution is to remove the target from the bibxml, as xml2rfc then inserts a correct target.

I disagree - the references should stand alone, and not count on side-effects of processing for semantic content.

[cabo]

I don't follow. Are you talking about references in published RFCXML (I'd agree) or what is served from bibxml (I don't agree)?

[stefanomunarini]

We retrieve the value for the target attribute from Crossref. The bibxml-service does not manipulate this value whatsoever. @cabo @rjsparks

[cabo]

Indeed, when I ask https://doi.org about

10.1109/access.2020.3012581

I get

URL: http://dx.doi.org/10.1109/access.2020.3012581

doilit simply ignores that value, relying on xml2rfc to fill it in based on doi_base_url, which defaults to https://doi.org/. My recommendation not to supply a target value based on this shady URL value stands.

[ajeanmahoney]

We retrieve the value for the target attribute from Crossref. The bibxml-service does not manipulate this value whatsoever.

Sounds like this issue should be opened at Crossref.

[rjsparks]

Yes, but we can do the right thing anyhow.

@cabo : Yes I meant what is returned from bib.ietf.org, not just what is published in RFCs. The fact that you are having to do target manipulation in your code (or that xml2rfc has to do so) is silly. Sure, that processing code can override whatever the bib service provides, but providing the value we would use is the stronger path to consistency.

What is your argument for leaving it off?

[cabo]

My main argument for leaving it off is to keep second-guessing to a single place. We already have the place in xml2rfc.

[rjsparks]

This would relieve xml2rfc or anything else from second guessing - instead it can just use what its given unless its explicitly told to override it from the command line.

[cabo]

I prefer to have it in xml2rfc as it seems the reaction times can be lower. But that is a preference, please do go ahead and fix it here, too.

[stefanomunarini]

@cabo @rjsparks can you confirm the fix should be done in bibxml-service?

[cabo]

@stefanomunarini I'm not the person to ask

[rjsparks]

Yes, please fix it in the bibxml-service.

[kesara]

359 was reverted.

361 is addressing this issue.