Friday, 1 May 2020

One space or two: has the great debate finally been resolved?

In my Effective Writing training course, I pose a provocative question. After a full stop, should there be one space or two?
It's provocative because there's no correct answer. Everyone will have been taught a particular way is correct, and many people get quite angry that not everyone else agrees.
I then present this XKCD cartoon as a talking point.
The intent is to illustrate that language rules - grammar, spelling, punctuation - are defined by consensus. What's correct English now is what people are actually writing and speaking now. That might be different to the consensus a hundred years ago, and, given today's level of connectivity, will likely be different again in a much shorter time period. But sometimes, we just don't agree on the rules, and that's where a style guide can be helpful - to help us agree to disagree, defer to an authoritative source, and move on.
Finally, I ask participants in the course to guess whether Microsoft Word's grammar check treats one space or two as correct. It's actually a trick question: Word can be configured to treat either, or neither, as being incorrect.
There is some objective history behind the question, and I have a personal opinion on what this means for the correct answer to the question. First, for most of the twentieth century, letters were typewritten and therefore monospaced, or fixed-width. That is to say, the typewriter carriage moved forwards a fixed amount when a letter was typed, no matter how wide the letter - i or l, m or w.
Text is easier to read if the space between sentences is clear and obvious. This means that a double-space after a full stop actually did make typewritten documents easier to read. I don't believe any of this is contentious. What is perhaps contentious is whether the same rule should still apply now that we all have word processors that can deal with letters of different widths.
Indeed, the width of the letter m is used as an actual unit of measurement in typesetting - called the em (as in the "em-dash", as opposed to an "en-dash").
My personal view is that it's bad form to change the content of data to make the formatting work properly. For that reason, I am content to say that one space is correct, and we should rely on our word processor (or web browser or whatever) to lay the text out neatly. Plus, it's very slightly less effort.
In fact, in HTML, multiple consecutive whitespaces condense down to a single space anyway. Type two spaces after a sentence, or anywhere else, and your browser will helpfully get rid of the second one.
All this is a rather long preamble to introduce a momentous piece of news: the world has decided - or at least, Microsoft has - that there is a definitive correct answer​ to this vexing question. And that answer is, as I would have hoped, that single spacing is correct. The linked article talks about the dubious research on the topic as well.
I don't think that this will settle the debate at all. Indeed, there will likely be some people who hold out against this decision precisely because it's Microsoft that's tried to settle it. But Microsoft's action tends to imply that the consensus has definitively shifted to the one-space view, and also ensures that most people will be encouraged to single-space in the future.
A version of this article was originally published on my company internal blog.

Friday, 3 April 2020

What's the plural of OS?

By virtue of my ever-popular Effective Writing training course, and despite no formal education in English above GCSE level, I am somewhat known for being a style and grammar authority / pedant. This means two things: first, my mother is very proud; and second, I get the occasional request to arbitrate on particular language usage.
This week: What's the plural of OS (where OS is the common abbreviation for Operating System)? Is it OSs? OSes? OS's?
I didn't know the answer offhand, so I looked it up in four different style guides and drew a blank. However, there are some good rules of thumb that might apply in this case.
One rule is: treat the abbreviation as though it were the expanded form. That would tend to favour "OSs", since we wouldn't add -es to the word System. However, this rule has some notable, and bitterly contested, exceptions, both in written and spoken English. For example, GIF is correctly pronounced jif, even though the initial G stands for Graphics.
Another rule is: be guided by pronunciation. This approach is very helpful to the reader as well as the author. I find this particularly helpful when trying to ascertain where to place an apostrophe in a word that naturally ends in an S. Is it pronounced St James' Place or St James's Place? This rule would favour "OSes" as the correct plural.
But there's one overriding rule in language, which is that common usage determines the rules. Dictionaries don't set or enforce language rules; they adapt to reflect the way the language is actually used. Even where a word or punctuation mark is technically wrong, the consensus of actual usage overrides it every time. The word "literally" is a very obvious case in point, having come to mean the exact opposite of its original definition. Language evolves, and every pedant in the world is powerless to stop it.
On these grounds, it appears that OSs is likely the more common, and therefore correct, usage. This article cites the Microsoft style guide and a mostly-respectful debate on the issue as evidence for this.
My final, and overarching, rule in a situation like this one is: as long as it's not obviously incorrect, choose an answer for a good reason, and then stick to it consistently. If a grammar pedant believes that "OSs" is inherently wrong, then you aren't ever going to change their mind; just as you aren't ever going to solve the war over spaces after a full stop.
Many thanks to my colleague Jack for raising this unexpectedly rather interesting question.
A version of this article was originally published on my company internal blog.

Thursday, 2 April 2020

The extraordinary, everyday technology that's making the lockdown bearable

Last night, a small group of my friends tried a virtual games night for the first time. The template was very helpfully set for me by a lunchtime social that a colleague ran last week. The key elements are:
  • Video conference so that we can see each other and the game screen.
  • Third-party hosted game that is already designed to be shared out onto participants' devices.
It was a very fun and relaxed evening, and a welcome diversion from being stuck at home indefinitely. What struck me afterwards was the sheer amount of computing power needed to make all this work. This could not have worked - or at least not as effectively - in any previous time period.
Because the game and the video conferencing platforms were both "in the cloud", it's impossible to say how much raw power they each provided. But I can say with certainty that we used a minimum of thirteen computing devices to power the games of five players, excluding all of the internet infrastructure that connected us together:
  • Each player played on their own phone (5).
  • Each household had a tablet or laptop to use for the video conferencing / screen sharing (3).
  • The game was hosted and screenshared from my home PC, but since I didn't fancy sitting at my desk all evening, I controlled it via remote desktop from an old laptop (2).
  • The game was supplied / licensed via Steam but required a third-party service ( to play (2).
  • Video conferencing and screensharing was via Webex [*], who are offering free personal accounts at the moment. (1)
It's easy to take all this for granted, but it's a fascinating reminder of the technological sophistication of the modern world.
Having tried it once, we'll definitely be doing it again soon, and will be looking to increase the complexity of the games as well. I'm happy to receive recommendations in the comments below :-)
[*] Other conferencing services are available; the most visible at the moment seems to be Zoom, but I am not confident that they have addressed all of their security and privacy faults yet.
A version of this article was originally published on my company internal blog.