Early-September updates: Library of Congress books, share links, Scopus and DOAJ marks

TL;DR: Since the late-August update: We’ve added queries to the Library of Congress book API. Every reference has a share link, journals show their Scopus and DOAJ standing, and the parser and the badges got another accuracy pass.

Books from the Library of Congress

When a reference is a book, RefRunner now first searches the Library of Congress (LC) catalog. LC catalogs the item in hand — all authors and editors, edition, place, publisher, year. This is a huge improvement, as the other repositories each have their own problems; Crossref and OpenAlex only have what a publisher deposited, with varying care—some publishers are still catching up to errors made in the past; Google Books often knows only the latest reissue; OpenLibrary catalogs what the Internet Archive knows about, so it’s broad, but uneven, and not always accurate. This is also a help for chapters in an edited volume—Crossref and OpenAlex often have these records for the chapter, but often are missing the actual details on the book. Whenever a book is cited, we now check LC for data, and use it unless there is reason to believe it’s wrong.

Share links

Every reference is now a URL you can send to someone. There’s now a share link next to the copy icon for the formatted reference:

Clicking that will give you a link like https://www.refrunner.com/cite/10.1080/10508406.2019.1570933?style=apa

You can also create your own link by just adding a DOI to the site URL like https://refrunner.com/10.1007/s11528-007-0040-x, or even refrunner.com/10.1007/s11528-007-0040-x.

But wait! There’s more! There’s also a new Share Link button:

If there are references without a DOI, a string is appended with the full text of the references just as they were in the paper. This lets you share any list of references with a (sometimes very long!) URL.

I used it this morning on my phone to build a reference list, get the link, and text it to someone with a dozen or so links. Though I don’t really recommend using RefRunner on your phone, it works pretty well!

Scopus and DOAJ marks

A reference whose journal is on Elsevier’s Scopus active source list shows a Scopus mark linking to its page; a journal in the Directory of Open Access Journals shows a DOAJ mark. A journal that has been removed from Scopus is flagged (only for work published after the delisting date), and the badge links to the journal’s JournalTrends page, which might explain why.

Lit Review mode and Batch Downloads

RefRunner will now make it possible to download PDFs of your reference list with a single click; there are a few steps that are more complicated than I’d like, but if you click the Lit Review button on the top toolbar, it’ll open up another toolbar like this:

The button on the left will open each PDF link in a new tab (after you give RefRunner permission to open popups). If you tweak your browser settings as described in the “PDF Settings” link at the right, you can have them all download without clicking to save each one. If you’ve configured your library’s proxy, the button on the right will appear and you’ll also be able to download PDFs through your library, again, conceivably with a single click. Use your judgement about how many articles you’re downloading at once. If you’re doing a large-scale literature review, you should check with your librarian. Large scale literature review projects (hundreds or thousands of articles) generally require special tools and permission from your librarian and publishers (the last such paper I helped with was recently accepted without revisions; if you need help, let me know!).

I’m preparing a talk to a group of doctoral students at the end of the month, so I should have some better notes on this soon!

Parsing and matching

Parsing citations and references in any of several styles is quite a trick. We’ve now checked over a thousand PDFs from multiple publishers as well as saved sessions from users. I keep thinking that things are “good enough,” but then another case appears that seems worth adding more code to handle. Here are a few recent fixes:

  • A huge problem is that sometimes multiple references will be combined into one. When that happens there hasn’t been much way for the user to resolve it. When I’ve found those examples, we’ve tweaked code to find ways to cover the various cases, but regular expressions can do only so much. Now, when two references are recognized as one (we look for signals like having two years), we present a “Split into two?” nudge. (You can even make it three!)

  • Corporate authors in AMA style (“World Health Organization. Title. …”) parse, paste, and format correctly.

  • References that end in a long URL no longer lose their title to it; AER-style lists (year last) parse; four-initial Springer authors parse. Seriously? Four initials?

  • We can now tell when a journal has renamed itself, and no longer call it a mismatch.

  • When the best search hit is rejected because it’s not a good enough match, the card now shows it as a closest match, not applied, so you can see what we almost used, which might indeed be what you were looking for.

  • A title ending in a question mark no longer swallows the publisher that follows it. The example was a book with a title ending with “…how should it be implemented? North River Press.” There isn’t any way a regular expression can tell whether “North River Press” is part of the title and the answer to that question, or the publisher (as it is in this case). That is resolved, at least for that book. :person_shrugging:

Badges

  • The mismatch badge reads “Best guess — verify”. The language before made it seem like the author might be to blame for the mismatch.
  • New: “Drops authors.” When a citation names authors the matched record doesn’t have — six names on a two-author paper — the card now says who, instead of quietly losing them in the diff. It’s the mirror of “Adds authors,” and both are the fingerprint of a reference list that was written from memory rather than copied from the paper.

Smaller things

  • Hover a row in the List view to see the full reference.
  • Keyboard shortcuts: As long as the cursor isn’t in a search box, type s l, s d, s c to Show the List, Diff, or Cards. You can also “t c” to Toggle the Cards.

As always: if RefRunner gets something wrong, hit the “RefRunner got this wrong?” button on the card, or message @pfaffman with as much info as you can provide. At a minimum the reference in question, but often the whole paper is very helpful