swcarpentry/modern-scientific-authoring

separate "Why not Word?" section in 01-mess.md

Opened this issue · 1 comments

You have Word vs. others sprinkled in throughout the text, but I think there would be significant value in having a whole section at the beginning taking on Word and not returning to it. Just focus on markdown / Latex / etc after that point. Here is more ammunition if you want to add a “Why not Microsoft Word?” section early:

a. track changes becomes cumbersome with multiple edits (you mention the version control limitation but it might help to explain it in Word-ese for those users. Also I have a Word doc in front of me with 10+ sets of track changes. It is illegible. I don’t know which changes were made on this last iteration).
b. insertion and placement of images and captions is time-consuming and is disrupted by text changes.
c. many changes have to be made manually instead of automated.
d. a specific example of automation is that reformatting text for different journals, etc., can require wholesale formatting changes.
e. The file can look different on different operating systems and Word versions.

Having failed to convert a few people myself, I agree that this is an important point. Word does have a merge tool that works reasonably well, though it can still be time consuming compared to a git merge even with no conflicts.

Comparing with any version

One strength of version control is comparing between any revision, not just the start and present, as @mandel01 mentioned. For example, journals may require a "diff" version for the reviewers. With Git/LaTeX, you can put a tag at each submission stage then use latexdiff (my own example). The Word equivalent would, I suppose, be to explicitly save each version along the way with a descriptive file name, then somehow remove all change history from that "milestone version" before moving onto the next revision stage.

Describing changes

When used properly, commit messages can show intent and summarize changes better than simply forcing the collaborator to read his or herself. You can revert a commit if it encapsulates a major idea that is now unwanted. Word equivalent? Reject all changes?

Validity

Where will the "true" file be kept, with the first author, everyone's email inbox, Dropbox? With, GitHub et al., forks are easily identifiable, and the Google Docs-esque services share a single file.

Figures

If you generate your figures with a script, you can put them in a directory and include them in your LaTeX/Markdown. If you regenerate the figs, the paper will be updated when it's recompiled. Word can do this, but it is not default, and when I tried it briefly I just didn't "trust" that the figures were actually re-included properly, since they can be embedded, linked, or some hybrid. If your plots are in Excel, they might do better at updating.

An anecdote on why not Word

One time, I was asked to help out with a grant proposal, so I accepted the shared folder to Dropbox. The "version control" scheme used was that each time someone edited a file, they incremented a number after their initials in the file name. Well, since there were a lot of parties involved, by the end the project directory took up 750 MB. Yes, almost a full GB of storage to produce a 50 page PDF.

I'm still not even fully convinced myself, but have not seen a path to working with Word documents for my own stuff. FWIW, I also don't consider Markdown as a viable option without automatic cross-reference numbering for figures and equations. Is that possible yet? Word is also still a pain for this as far as I know.