How to publish a working paper on Arxiv?

It is a very good practice to publish working papers on Arxiv. Here are some tips on how to do this.

Why publishing working papers on Arxiv? There are many good reasons (see this, that, and other):

• being read and cited early
• being discussed in the general press (see this example)
• being able to claim precedence in case of competing research on hot topics

When publishing on Arxiv? Either when the paper is in a good shape, and has a chance to be read and cited, or after the paper submission. In the context of double-blind peer-review, it is a good idea to wait a little bit, so that there is no obvious timing correspondence between submission and dissemination through pre-publication (check the conference/journal guidelines).

Checklist before submitting a working paper on Arxiv

• remove all “under submission”, “submitted to”, “journal name”, “conference name”
• identify the appropriate Arxiv categories (eg “CS.SE”)
• if appropriate, de-anonymize it: add authors, restore hidden URLs and content
• add an “Acknowledgement” section, to thank all the people who gave you feedback
• force Arxiv to only create a PDF version (and not a DVI one), add \pdfoutput=1 at the beginning of the document

Tool support for submitting to Arxiv

• https://github.com/dougalsutherland/arxiv-collector
• https://github.com/google-research/arxiv-latex-cleaner​

How to submit to Arxiv without uploading the Latex sources? There are plenty of solutions:

• Trick the latex compiler (pdflatex trick; lualatex trick)
• Create an overlay latex document, https://mssun.me/blog/how-to-bypass-arxiv-latex-generated-pdf-detection-in-six-lines.html (Does not work anymore see https://gist.github.com/monperrus/2852987a9a8340fc19ae6bb55a76f62e)

What to do next? A paper on Arxiv is in most of cases invisible. There are too many papers on Arxiv (and outside Arxiv as well) for people to notice it or take the time to read it. Once the paper is on Arxiv, the next step is to let appropriate people know about it. The persons most likely to read and like it are those working in the same area. The next step is thus to send a personal email to those 10-20 authors having worked on the most closely related papers.

See also:

• Open Access Publishing with arXiv, by Tommy Ohlsson, KTH (KTH is a contributing organization to Arxiv)
• arXiv.org - Frequently Asked Questions
• Popularity of arXiv.org within Computer Science, by Charles Sutton, Linan Gong
• Making unblinding manageable: Towards reconciling prepublication and double-blind review by Charles Sutton
• PLDI Double-blind FAQ
• The impact of Arxiv CS.SE according to Google Scholar