Feature request: auto recognising arxiv pdfs

[alexei]

Hi,

most of the current research papers I read are on the preprint server arXiv.org (or derivatives like scirate.com) and this must be true of many people as it's the default repository for new papers in many fields. Bookends does a great job identifying pdf's with a DOI number embedded but can this be extended to an arXiv number if the DOI and ISBN etc fail? At the moment each preprint reference has to be linked by hand to metadata by copy-pasting the arXiv number in the the autocomplete screen with the arXiv filter.

[Jon]

Please see the release notes for 13.0.3.

http://www.sonnysoftware.com/updates/updatehistory.html

In particular

• Automatically import arXiv PDFs when you import the references's metadata

So if you do an arXiv search you get the PDF when you import it. You're asking for the opposite. Whether that's do-able depends on how easy it is to unambiguously dated the arXiv number. You can send me a few example PDFs and I'll look.

Jon
Sonny Software

[alexei]

Hi Jon, thanks for the quick reply! Yes, I was thinking in the other direction, given the pdf automatically lookup the metadata (after landing in the watch folder for instance). Sample pdfs are readily available from https://scirate.com or https://arxiv.org/list/quant-ph/new . They will typically have an arXiv number in the format "arXiv:YYMM.XXXXX" e.g. arXiv:1802.03540 on the first page. More on the format here: https://arxiv.org/help/arxiv_identifier

[Jon]

Thanks, I'll take a look at it tomorrow.

Jon
Sonny Software

[alexei]

Hi Jon, the arxiv import is really great and so much better than manual matching.

Is it possible to change a detail of the import? Bookends seems to take the Journal-ref field as the authoritative reference data. Having this data is great but the arxiv version and the final published version of a manuscript are different entities. There is no requirement on authors to upload the final version of a paper and it may change substantially after going through review. At the moment references end up looking like they're the authoritative source but the pdf is a preprint and the metadata was entered by hand by the authors.

Could we have arxiv papers always imported as arxiv papers (with a link to the published version as a note/url)? This would keep it clear which version you have in your library.

[Jon]

Sorry, I don't understand what you want. If you want some different field imported as (what? the journal name, date, volume?) please send me examples (off-forum) of PDFs (I assume you're talking about importing from PDFs, but maybe you mean direct searches).

Jon
Sonny Software