From 1319ebb828071873efbf331ca17a25813381c794 Mon Sep 17 00:00:00 2001
From: Fabien Benureau <fabien.benureau+git@gmail.com>
Date: Sun, 29 Mar 2015 03:45:04 +0200
Subject: [PATCH] added some doc in readme

---
 readme.md | 29 +++++++++++++++++++++++++----
 1 file changed, 25 insertions(+), 4 deletions(-)

diff --git a/readme.md b/readme.md
index 5ff4cf5..ca6a13c 100644
--- a/readme.md
+++ b/readme.md
@@ -24,21 +24,42 @@ Import existing data from bibtex (pubs will try to automatically copy documents
 
     pubs import path/to/collection.bib
 
-or for bibtex containing a single file:
+or for a .bib file containing a single reference:
 
     pubs add reference.bib -d article.pdf
 
-you can also retrieve the bibtex from doi.org by giving the DOI:
+pubs can also automatically retrieve the bibtex from a doi:
 
     pubs add -D 10.1007/s00422-012-0514-6
 
-The pdf must still be downloaded manually.
+or an ISBN (dashes are ignored):
+
+    pubs add -I 978-0822324669
+
+The pdfs must still be downloaded manually.
+
+
+References always up-to-date
+----------------------------
+If you use latex, you can automatize references, by creating a bash script with:
+
+    #!/bin/bash
+    pubs export > references.bib
+    latex manuscript.tex
+    bibtex manuscript
+    latex manuscript.tex
+
+This ensure that your reference file is always up-to-date; you can cite a paper in your manuscript a soon as you add it in bibtex. This means that if you have, for instance, a doi on a webpage, you only need to do:
+
+    pubs add -D 10.1007/s00422-012-0514-6
+
+and then add `\cite{Loeb_2012}` in your manuscript. After running the bash script, the citation will correctly appear in your compiled pdf.
 
 Requirements
 ------------
 - python >= 2.6
 - [dateutil](http://labix.org/python-dateutil)
-- [pyYaml](http://pyyaml.org)
+- [pyYaml](http://pyyaml.org) (will be deprecated soon)
 - [bibtexparser](https://github.com/sciunto/python-bibtexparser) >= 0.5.3