To handle complexity and versions, I do the following when I start writing a paper in LaTeX.
Some things to note:- Create new directory Conf1.
- Within this, create a new file Conf1.tex, and two new directories: tex and images.
- Conf1.tex contains: document class, title, author, bibliography style, etc., which are specific to the conference/journal. This is followed by \input{tex/main}.
- Create main.tex in directory tex. The tex directory contains all the LaTeX code, and could have further subdirectories if required (e.g., I usually have a directory for Experiments, another for Algorithm, etc.). All tex files are included in the paper by adding \input{tex/subdir/abc.tex} to main.tex in the appropriate place.
- All package inclusions (\usepackage{}), command declarations, etc. required for the paper are included in tex/include.tex.
- The first line of main.tex is \input{tex/include} followed by \begin{document} and \maketitle. Include all your LaTeX code after this. End with \bibliography{tex/biblio} followed by \end{document}, where tex/biblio.bib is the bibliography file.
- All images are stored in the directory images.
- You can use git/svn to do versioning. Until I get used to them, I am using the following method. When I have a significant frozen version, I copy the directory tex to texN where N is 1,2,3.... The timeline of versions from the earliest to the current is tex1, tex2, ..., texN, tex.
- I put the .bib file into tex because I found the bibliography changes as the paper develops.
- On the other hand, images usually accumulate. So, only the directory tex needs versioning.
- In order to use the material for another conference/journal, follow the steps above to create Conf2 and Conf2.tex. Then copy images and tex from Conf1 and compile. Your first draft of Conf2.pdf is ready!
- I read somewhere that using PDFLaTeX gives more compact files, also avoids some font problems which arise due to DVI to PDF conversion when using LaTeX. I also had problems using Beamer with LaTeX.
- If you decide to use PDFLaTeX, remember that it does NOT accept images in .eps format. I found that .png files did not render text well on conversion. Creating / converting all images in / to .pdf format (using Inkscape) worked for me.
- I use Inkscape for most drawings, and Gimp for screen capture (and also for format conversion).
No comments:
Post a Comment