As I’ve gone through UTP and started modernizing,
and finding (or all too often, recreating) lost utilities,
I noticed several quirks that bothered me—overuse
of “will” and “simply,” and lots and lots of passive voice.
I fixed those that I noticed as I went through the book,
but obviously missed a lot more.
I finally decided to take a more systemic approach. Maybe
it was when I was going through the regular expressions
section that I thought of it:
$ grep -En '(is|are|was|were)( not)? [a-z]+ed’ *.t | wc -l
1057
$
This is in the “2.0 (UTP Revisited)” branch. As the main
branch returns 1348, I’ve made a dent. The thing is,
most lines matched contain more than one match;
VScodium says there are 2090 matches total.
Of course, this isn’t the total number of passive constructs:
this doesn’t catch any patterns with a line break, and there
are other passive constructs that I’ve run across. But this
is a start. Difficulty ranges from deleting a word or two, to
moving a few things around, to completely rewriting the entire
sentence. Often, you’ll notice the active construct is one or two
words shorter than the original; other times, (especially with
the rewrites) it might be a few words longer. I didn’t think to
do an initial word count, and it doesn’t matter anyway since
we’re adding new material (including a new chapter on using
Markdown and XML with *roff).
Anyway, is anyone interested in picking up one or more files
to scan and fix? (Feel free to comment on my modernization
attempts as well.) I’d started with app[a-c].t, so I’ll finish them
up and tackle any chapters left unclaimed.
Thanks much!
— Larry
