Improvement Requests for MODS XML Import
Thanks for the UnAPI support in the developer build--this will be useful. It is currently capturing less bibliographic info than the COinS translator was for refbase-0.9.0, though.
This brings up a few notes:
* Can extent work for the "page" type in addition to "pages"? The schema doesn't require either, but does suggest "pages" (which Zotero uses), but they give examples using "page" & other sources follow this convention.
* Can the identifiers be imported from the main MODS entity, in addition to from the relateditem "host" branch? The schema states identifiers can be in the root of a MODS record & it is unclear to me why a DOI should be associated with the host journal, rather than with the article.
* It'd be nice to import the "abstract" (and not just "notes")
This brings up a few notes:
* Can extent work for the "page" type in addition to "pages"? The schema doesn't require either, but does suggest "pages" (which Zotero uses), but they give examples using "page" & other sources follow this convention.
* Can the identifiers be imported from the main MODS entity, in addition to from the relateditem "host" branch? The schema states identifiers can be in the root of a MODS record & it is unclear to me why a DOI should be associated with the host journal, rather than with the article.
* It'd be nice to import the "abstract" (and not just "notes")
We will probably keep the changes to "pages," but I don't know if we'll keep the location of the DOI in the host or not. It doesn't seem to break anything yet, but it still seems weird to associate a single DOI with a journal (as if it were equivalent to the ISSN).
I haven't changed our "abstract" into a "note" yet--that breaks other software which treats the "abstract" specially.
Perhaps something like:
<physicalDescription>
<internetMediaType>application/pdf</internetMediaType>
</physicalDescription>
<location>
<url displayLabel="PDF">http://arc.nucapt.northwestern.edu/refbase/files/Karnesky-2007_9601.pdf</url>
</location>
I'd still be interested in a method of linking a PDF to a record such that Zotero could auto-download it.
<location>
<url displayLabel="Electronic full text" access="raw object">http://some/url/to_a_file.pdf</url>
</location>
It'd be nice for zotero to auto-download the full text if it sees something like this.