search

ropensci / RNeXML

Implementing semantically rich NeXML I/O in R
https://docs.ropensci.org/RNeXML
Other
13 stars 9 forks source link

Automatically add value of $USER to document-level metadata (dc:creator) #86

Closed cboettig closed 9 years ago

cboettig commented 10 years ago

Rutger does this in the PERL version, seems like it could be handy.

I do worry a bit that there's a tendency not use any hierarchical organization of metadata (see #23) that things can get confusing. For instance, when associating publications with the NeXML they are currently just added as annotations at the document-level as well. (e.g. https://github.com/ropensci/RNeXML/blob/master/inst/examples/geospiza.xml)

Perhaps I should be more aggressive about using hierarchical / nested meta elements, though it raises the additional challenges discussed in #23.

rvosa commented 10 years ago

Along similar lines, a timestamp is probably nice to have as well. Though then the question becomes: what does a program have to do if it reads a file with a timestamp and writes it again: append another timestamp? Overwrite it?

cboettig commented 10 years ago

Good question. Wondering if there's a good way to coordinate thinking about this kind of thing at a higher level so at least whatever we do is consistent across parsers and with most users expectations?

cboettig commented 9 years ago

A top-level timestamp is added only if one does not yet exist by add_basic_metadata since d93056b2496f951896ef35453e0115cafe9525cf, USER added in a2c94a4b613999d7fa6604cd856294be548f1e88 . Currently this leaves updating the timestamp up to the use of more manual controls by the user. (automatically appending a new timestamp each time seems messy without richer metadata describing the differences, and changing existing metadata without notice seemed rude).