Improve XMP

[koppor]

Follow up of #3895

• Why does the order of XML elements change at a second export? This IMHO should not happen.
  
• In non-BibTeX fields, There should a be a cleanup made. latex2unicode at each field.
  "Proceedings of the 10\textsuperscript{th} Central European Workshop on Services and their Composition ({ZEUS} 2018" should become "Proceedings of the 10th Central European Workshop on Services and their Composition ({ZEUS} 2018)"
• In non-BibTeX fields, a latex cleanup should be executed. - Maybe this is done by the latex2unicode thing: "{CEUR} Workshop Proceeding" <-- Please no {} in the XMP.
• I am not sure whether <rdf:li>bibtex/url/https://adr.github.io/madr/</rdf:li> is the correct RDF searlization for an URL. According to https://www.w3.org/TR/rdf-syntax-grammar/#section-Syntax-list-elements, it should be <rdf:li rdf:resource="https://adr.github.io/madr"/> - this affects export and import
• Keywords should be saved differently

      <dc:subject>
        <rdf:Bag>
          <rdf:li>ADR</rdf:li>
          <rdf:li>MADR</rdf:li>
          <rdf:li>architecture decision records</rdf:li>
          <rdf:li>architectural decision records</rdf:li>
          <rdf:li>Nygard</rdf:li>
        </rdf:Bag>
      </dc:subject>
  

  becomes ; ADR; MADR; architecture decision records; architectural decision records; Nygard, which is good enough. (See Ctrl+D in Acrobat Reader)

Test entry:

@InProceedings{KoppAZ-MADR-ZEUS-2018,
  author    = {Kopp, Oliver and Armbruster, Anita and Zimmermann, Olaf},
  title     = {Markdown Architectural Decision Records: Format and Tool Support},
  booktitle = {Proceedings of the 10\textsuperscript{th} Central European Workshop on Services and their Composition ({ZEUS} 2018)},
  year      = {2018},
  volume    = {2072},
  series    = {{CEUR} Workshop Proceedings},
  pages     = {55--62},
  publisher = {CEUR-WS.org},
  keywords  = {ADR; MADR; architecture decision records; architectural decision records; Nygard},
  url       = {https://adr.github.io/madr/},
}

Test setting now at https://github.com/JabRef/jabref/tree/feature/xmp-provement/src/test/resources/pdfs/KoppAZ-MADR-ZEUS-2018

• Fix XREF Error when writing XMP Metadata to PDF #7540

[johannes-manner]

Point 4: url rdf serialization. I am also unsure how to serialize urls right.. I found no clear description, but I found an example on the dublin core home page (point 4.15) http://www.dublincore.org/documents/2000/07/16/usageguide/generic/#rights
It is a hint, that serializing the url in the way we do in dublin core is ok. What's your opinion?

Point 5: "Keywords should be saved differently" - why to save the keywords differently? All other fields are also stored as lists in the xmp metadata. Maybe the interoperability to other schemas is influenced by a single item list with a custom separator.

[koppor]

Point 4: I thought about seeAlso (see https://www.w3.org/wiki/UsingSeeAlso), but it seems to be more complicated. See that bibtex/url can only be read by JabRef and not by other tools. Storing XMP should be interoperable with other tools, too.

Point 5: Keywords should appear as follows:

With the current export:

So, other tools will have difficulties to parse the information correctly.

[koppor]

Another library for XMP: https://github.com/drewnoakes/metadata-extractor

[Marathon2112]

I'm linking #7925 here as well.

[koppor]

This task needs a proper requirements engineering, test cases and trial (because of the other available library). One has to test with Acrobat Reader and other PDF tools.

[Siedlerchr]

Refs #8789

[Priyanshi-1001]

@koppor I tried to reproduce this Error,
Added Test Entry in Entitles and tried to export XMP meta data to PDF but its not succeeded, Am I Doing correct?

[koppor]

@Priyanshi-1001 I don't understand. You should provide detailed steps what you are doing. Do you use the test PDF linked at https://github.com/JabRef/jabref/tree/feature/xmp-provement/src/test/resources/pdfs/KoppAZ-MADR-ZEUS-2018?rgh-link-date=2018-04-12T15%3A02%3A56Z? What do you see in your PDF program? What do you mean by "not succeeded"?