On Fri, 26 Dec 2014 23:27:37 -0500, Nick Dokos <ndo...@gmail.com> wrote:
Anyway, color me deeply suspicious of the "study".
Indeed!
The study touches only a few of the inherent difficulties in document
production. Its major flaw is that it draws any conclusions at all
recommending that authors produce documents one way or another. Personally
I am always disappointed when someone requests a document in MS Word
format, because that means I'll have to fire up Libre Office and shove my
text through it, rather than using whatever other system I happen to have
been using. I do not believe that I currently own a system with genuine MS
Word.
As well as having insufficient control of variables, and a flawed
understanding of what is involved in "document preparation," the study
also has a marginally small sample size. Any study for any purpose that
presents "statistics" with sample sizes smaller than 30 is immediately
suspect. I won't even begin to address the misinterpretation of
correlation as causation that appears in the "softer" sciences, nor their
necessity for sample sizes far larger than 100, nor the tendency in some
fields to mistake a time series as a set of samples.
MS Word works extremely well for "one-off" small papers. Little investment
of effort is required for a naive person to produce adequate results, and
as every user of emacs knows, that's pretty much the opposite of emacs.
On the other hand, MS Word has historically been a terrible tool for
producing large documents, or documents that are to be maintained by a
group of people, or over several years or decades. Handling Word's "Master
Document" provision without being crippled by corrupted documents is an
art form unto itself. The standard advice among experienced users of Word
has always been, "Don't Use Master Documents!" When a group of people are
all editing versions of a document, any attempt to use standard formatting
in Word requires substantial effort to prevent naive contributers from
reformatting outside the established styles, or even breaking all the
styles. Furthermore, Word documents are in general not amenable to
incremental version control as commonly used by coding teams.
My conclusions? If your paper is trivial and you are under pressure to
produce it quickly, then MS Word might be the best tool. Established
journals should attempt to allow contributions in more than one format,
and restriction to MS Word format is a bad idea, no matter how much some
people like the apparent ease-of-use that MS Word provides. Attempting to
extend the "study" to include org mode would be a waste of effort.