When preparing a final version of a figure, it appears that some changes could be more easily made in a graphics software (gimp, inkscape, scribus, adobe) as opposed to R. Such changes might be done more efficiently, and more to the publishers standards, by the publisher. Apparently, the publisher uses some such software to makes final changes to the image anyway. Presumably, these changes can do it more efficiently and to their standards. Also, some tasks would be trivial or would be done anyway by the publisher.
Some tasks that I am currently doing to prepare figures for a journal article:
- move / add text
- change font
- change line thickness
- change background color
Something that would be neat to do:
integrate figures into text, in the style of Tufte:
- What formatting is commonly done by a publisher rather than an author?
- Is it reasonable to make requests?
- If so, what work will a publisher be willing to do?
In many particle physics journals the answer is they do nothing. You deliver a latex source file that uses their class and figures that meet their standards or they send it back.
And in all truth, I like it that way.