19 Comments

  1. Something I’ve noticed with this, and other options for automatic Bibtex generators (e.g. Zotero’s export feature) is that they don’t handle special characters at all well. For instance if an author has an accent in their name, like “Joe Bløggs”, the Tex engine doesn’t understand the special character and spews out garbage requiring me to manually tidy up my Bibtex file by escaping the special character. It would be great to have the option for the exporter to escape the special characters by generating the author’s name as (in this case) “Joe Bl\o ggs” or “Joe Bl\'{o}ggs” or whatever, depending on the character required.

    • tom

      Hi Ian,

      Thank you for the feedback. I agree, special characters are a problem in BibTeX formatted citations. Currently, our service replaces the most common special characters, including greek letters, mathematical operators and a subset of all possible accents (in total ~120/900 characters). We are already working on a better solution which will be deployed in the near future. Your suggestion to let the user decide might in fact be the right way to go. I will add it to the list of feature requests.

      Thanks again!
      Tom

  2. I have tested several of my papers and all cases conference papers are incorrectly returned as journal articles (@journal) when they should be conferences (@INPROCEEDINGS).

    The examples tested:

    10.1109/AFRCON.2011.6072062
    10.1109/ICTAI.2009.35
    10.1109/IECON.2013.6699120

    On the contrary, this one was correctly detected and provide the relevant information.

    10.1109/TII.2012.2219063

    I will fill a bug in github.

  3. Karl at CrossRef here. This is a problem with how we are generating bibtex. I’ve been using the Citation Style Language and their bibtex style definition which does not always produce accurate or valid bibtex.

    I’m now in the process of rewriting our bibtex output to use a different library to produce the bibtex itself. I’m hoping that this will address encoding issues as well as type accuracy.

  4. Amir Sariri

    Thank you for this wonderful post and helpful comments. I work on knowledge flow and the data I use for my current research is a set of citation data from hundreds of focal academic papers (backward citations), which will eventually end up thousands of observations.

    It obviously can’t be done by hand and we are writing a program that takes the name of the article from reference section of each focal paper, searches it in Google Scholar, and collects the BibTex file.

    However, I am hearing too much about the erroneous/irrelevant information in Google Scholar BibTex codes. Does Mendeley maintain a better collection than Google Scholar, and thus does it make sense to write our program for Mendeley (or maybe another service provider) instead? I mean I don’t really have any loyalty to Google, I’d want to collect the data with the least noise.

    Any ideas?

    Thank you in advance.

Leave a Reply to Ian GodfreyCancel reply