[Arcanum-XIII]

It seems there is some confusion for the author between the display part and the on disk part. text will not be saved with the ligature, it will still be pure unadorned text when saved... and for the confusion that could arise between a simple quote and a typographic one, it's not coming from the font but from the editor rendering engine (word does those kind of change, IntelliJ does not for example, but both can display correctly ligature)

[yorwba]

I'm sure the author is well aware of the difference between encoding and display, since they have apparently designed their own fonts.

However, their argument seems to be that since ligatures are just a hack to make plain text look like something fancy, you end up with confusing interactions due to the lack of context.

Now I don't think this is likely to cause lots of problems in practice, since most people won't use both ligatured => and Unicode ⇒ in the same Haskell code; but in principle the potential for confusion is there. Imagine if some clever font designer made ' look like ` when between a space and a letter, the font would be unusable in any language where that is an important distinction.

Personally, I'm not going to use ligatures for coding, firstly because I'm accustomed to the "normal" look and secondly because fcitx makes it really easy to type arbitrary Unicode when I want it.

[rossy]

I'm aware that Butterick is the author of several fonts, but this article has a surprising amount of vitriol from someone who should know that the choice of a programming font doesn't affect anyone but the programmer who chose it. During the article, he takes several opportunities to point out that his opinions are actually facts.

Typographers usually have a lot of responsibility because their decisions can affect the readability of text for many other people, and this is why sage advice from experienced typographers is usually very handy, but in this case, he's complaining about something that each individual developer has to manually opt into. This is outside the realm of normal typography that he has authority in. The editor settings of an individual developer are extremely personal. Developers don't follow trends just for the sake of it. They set up their environment in the way they believe is most productive for them, because they know only their output matters. If an individual developer believes they are more productive with Fira Code, that's nothing a typographer should lose sleep over. If I believe that I'm ten times more productive when I code in Papyrus, that's no one's business but my own.

[yorwba]

> Developers don't follow trends just for the sake of it.

I disagree with this. The problem with lots of programming tools is the difficulty of judging which options are the best for your use case, before you have the experience of using them for a long time. So there is a real temptation to just go along with what everyone else seems to be doing; since they presumably know the benefits. There doesn't even have to be a real trend; it's enough when there's one in your perception.

I have certainly done quite a lot of trend-following myself. I have used relative line numbering in vim ever since I started using it, simply because the first person I ever saw using vim had that option active. Later, when I found myself doing mental arithmetic to get absolute line numbers, I was simply too lazy to make the switch. I have actually only just now gotten around to changing my config.

So someone who is thinking about switching to a ligatured code font might benefit from an article like this, if it turns out that the issues raised would be too annoying for them. If they don't think those will be a problem, however, they can just go right ahead and make the switch.