Tuesday, January 15, 2019

Super-picky Writing Advice

There are patterns to bad writing. I'll give some examples based on a blog post I was sent. It's also based -- indirectly -- on some of proposals I saw for PyCon and PyDataDC.

For the conference calls for papers, I can ask a few questions of the author, but that's about it.

For the blog post, I suggested a bunch of changes.

They balked.

Why ask for advice and then refuse to do anything?  (We can conjecture they wanted a "good job" pat on the head. They didn't want to actually have me give them a list of errors to fix.)

One of the points of contention was "Not everyone has your depth of expertise."


The blog post was on Ubuntu admin: something I know approximately nothing about. Let me step away from being an expert, while sticking closely with being able to write. I work with editors who -- similarly -- can write without being deep technology experts. I'm trying to learn what and how they do it.

In this case, my editing was based on general patterns of weak writing.

  1. Contradictions
  2. Redundancy.
  3. Waffling.
  4. Special-Casing.

Any blog post on Ubuntu admin that starts with "This is not just another blog post..." has started off with a flat-out contradiction. It emphatically is another blog post. You can't rise above the background noise of blog posts by writing a blog post that claims it's not a blog post. Sheesh. Find a better hook.

Any blog post on Ubuntu admin that includes "This blog post assumes the reader is familiar with linux sys administration." As if -- somehow -- a reader interested in Ubuntu admin could be confused by the required skills. It's clearly redundant. Cut it.

(This led to an immense back-and-forth with repeated insistence that *somehow* someone once got confused and *something* bad happened to someone. Once. My response was adamant. "It's redundant. The title says Ubuntu. That covered it. More repetition is an insult to the reader.")

Anything that has "this may nor may not work depending on your filesystem" is flat-out confusing: it covers both bases. Does it work or does it not work? Which is it? Clearly, there's some kind of precondition -- "must be this file system" or "most not be this file system" -- buried under "may or may not work." It's not that I know anything about Ubuntu file systems. But I can spot waffling.

Indeed, when you look at it, this is a "hook" to make the blog post useful and interesting. Some advice doesn't work. This advice always works. Simple statements of fact are better than contradictions and waffling.

Finally, there was a cautionary note that replacing "/swapfile" with "/ swapfile" would brick your OS. Which. Was. Crazy. It's really difficult to arbitrarily introduce spaces into shell commands and still have proper syntax. Sometimes a shell command with rando spaces may have proper syntax and may work. Most commands won't work at all with a rando space added. Try some and see.

What's more important is only one random punctuation error was listed. The special-case nature of this was a tip-off that something was not right with this advice.

What about a random ">" in the command? Or a random "|"? Not covered. A single space was considered worthy of mention. The rest. Meh.

(If you want details, it's was `dd -of=/swapfile ...`. If asked, I'd guess `dd -of=/ swapfile` is a syntax error because `swapfile` isn't a valid operand to `dd`. Not an expert. I didn't check. AFAIK, the author didn't actually check, either.)

None of this editing was based on any vast expertise. It was simple editorial work looking for some common problems with hasty writing.

  1. Avoid clear contractions.
  2. Avoid redundancy.
  3. Don't waffle.
  4. Special cases are instances of a more general pattern. The pattern is more important than the case.

One of the contentious back-and-forth issues was "I'm not writing a book, it's only a blog post."

Wait. What? Blog posts are often more widely-read than books. Look at Stack Overflow. If one of my books had the kind of readership my Stack Overflow answers have, I'd be living on royalty payments alone.

A blog post requires the same depth of care as a book.

This applies to proposed talks at conferences, also. The proposal needs to be carefully edited to reflect the final presentation's care and quality.

Thank goodness, most of the 100's of proposals I've looked at rarely have the four obvious problems listed above.

The problems I see in conference proposals are minor.

  1. Incompleteness. A 45 minute talk boiled down to 4 bullet points doesn't give us any confidence. It's hard to imagine filling the whole time with useful content when looking at a four-sentence outline. Will it be rambling digressions? Or will we have disgruntled attendees who had hoped for more?
  2. Weirdly cute style. Things like "This is where I jokingly outline something something and the real fun begins." We assume everyone is witty and charming, you don't need to tell us. We assume all talks will be fun. Can we move on to the Python (or PyData) topics you'll cover?
  3. Sales Pitches. "[Speaker's name] is a respected industry expert who delivers exciting and transformative keynote addresses and will dynamically cover the state-of-the-art blah blah blah..." Please. What Python topic is this? We like to review the outlines without speaker information; we need to focus on the content. Subverting this by including the speaker's name in the description or outline is irritating.
I'm really pleased we see very few PyCon Code-of-Conduct problems in the calls for proposals. That is a delight.

Editing is hard work. I'm sorry to report that editing means making changes. If you ask for editorial advice, it helps to listen.

No comments:

Post a Comment