Propagate metadata fixes to Crossref?

Raised in #5701: Apparently the DOI record contains metadata that we have not been updating when we process metadata fixes.

I haven't investigated but this sounds like something we should be doing in the bulk metadata update script if possible.

We should also consider recording in the XML when a change was made to a paper's metadata (e.g., an `updated=""` attribute). This would allow us to see which metadata needs some downstream bulk processing e.g. in a Crossref update. (We should consider whether there are any author-page updates that could trigger a Crossref update for a paper, such as adding an ORCID iD.)

From https://en.wikipedia.org/wiki/Crossref:

> When a scholarly journal publishes an article, typically the publisher will enter the following information about the article into Crossref: journal name, article DOI, publication date, journal volume, issue, and page, URL of article as well as journal, and number of pages. Optional metadata that can be entered includes the text of the article abstract, [ORCID](https://en.wikipedia.org/wiki/ORCID) iDs of the authors, funding information, including funder registry IDs and funding award numbers, license information, and similarity check URLs.

No file or test is named. Start by locating the bulk metadata update script and tracing how metadata fixes reach generated XML and Crossref-related data; clarify whether author-page changes such as ORCID updates are in scope. Done should define the affected metadata, record the relevant XML change information, and verify that required Crossref updates are propagated.

由索引模型根据 Issue 内容生成。