Most of the digital version of published papers use Postscript or PDF as their document format, which helps to ensure that the document looks as intended on various platforms and in print. One of its drawbacks is that text is just a large array of glyphs that are to be placed on a page, semantics such as text being a heading, caption, text in a figure, or a reference, is lost.
This causes problems when you want to make automated use of a paper, e.g., to extract the references inside, enable fulltext search, or generate bibtex entries. For example, for inclusion into the ACM Digital Library you have to provide the LaTeX source and bibtex file to help them get the proper metadata in the first place. Google Scholar or RG seem to invest lots of effort to find out which papers are yours, who is cited by whom, etc.
There was much research on topics like the semantic web, are there already solutions for this problem that could be used for research documents at scale? To improve the situation, why not have a new data format (or just include this metadata in PDF) and make this information mandatory when publishing it?
There is no need to "invent" something new for this. PDF already supports metadata by default.
There are the basic PDF metadata that are not really much but good for beginning.
And then there is also XMP, the metadata library of Adobe that support adding RDF data to all objects in a PDF. The Downside of this is that you need to add all the metadata by hand in Acrobat or other (in most cases commercial) tools.
Some more Integration to the editing process would be the most helpful thing for the "world of metadata".
http://www.pdfa.org/2011/09/pdfa-metadaten-xmp-rdf-dublin-core/
There is no need to "invent" something new for this. PDF already supports metadata by default.
There are the basic PDF metadata that are not really much but good for beginning.
And then there is also XMP, the metadata library of Adobe that support adding RDF data to all objects in a PDF. The Downside of this is that you need to add all the metadata by hand in Acrobat or other (in most cases commercial) tools.
Some more Integration to the editing process would be the most helpful thing for the "world of metadata".
http://www.pdfa.org/2011/09/pdfa-metadaten-xmp-rdf-dublin-core/
I agree that it is easy to add metadata to PDFs (including putting everything into the filename), my intention is to have a standardized format that papers should follow so that you can get for example a reliable bibtex file when you have the PDF instead of searching the web for it. At the moment, the document and its metadata are at different places (ACM DL, Springer Link, DOI, ...) so there is always indirection.
I think latex pdf rendering makes complete pdf. however all journals hv their own pdf conversion dervers. we can just hope them to support metadata.
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering