Do your writing in text files accompanying the image files (figures). The LaTeX code is just instructions for how to render the various text sources arrange the figures on pages to be printed or rendered as slides.
It separates the flowing creative experience of writing and documenting what happened in the experiment from the fiddling creative experience of rendering, editing, and presentation to ensure that the text and figures line up appropriately and are on appropriate pages.
Separating fact finding from presentation is an important barrier in the scientific method.
Yup. When I rerun my things, in latex I just overwrite the plots file (pdf/png) and compile latex. In word I have to find where it was and replace it there. It’s way easier on latex if you make your code just write plot files in the same location.
You can use a symlink to point to the figures directory of a certain run of the code. Add git history to the mix, and now you have an auditable record of what version of the code’s output ended up in each version of the paper.
And the ease with which you can generate hundreds of lines of page with a simple text template and code.
Right.
Do your writing in text files accompanying the image files (figures). The LaTeX code is just instructions for how to render the various text sources arrange the figures on pages to be printed or rendered as slides.
It separates the flowing creative experience of writing and documenting what happened in the experiment from the fiddling creative experience of rendering, editing, and presentation to ensure that the text and figures line up appropriately and are on appropriate pages.
Separating fact finding from presentation is an important barrier in the scientific method.
Yup. When I rerun my things, in latex I just overwrite the plots file (pdf/png) and compile latex. In word I have to find where it was and replace it there. It’s way easier on latex if you make your code just write plot files in the same location.
You can use a symlink to point to the
figuresdirectory of a certain run of the code. Add git history to the mix, and now you have an auditable record of what version of the code’s output ended up in each version of the paper.You can be so anal and precise about everything.