Read and understand code faster with programming ligatures in Fira Code font(visualstudioextensions.vlasovstudio.com)
visualstudioextensions.vlasovstudio.com
Read and understand code faster with programming ligatures in Fira Code font
https://visualstudioextensions.vlasovstudio.com/2017/05/01/improve-code-readability-with-programming-ligatures-in-fira-code-font/
42 comments
I had every one of those concerns too. In spades. And then I actually tried using it for real, and none of the concerns turned out to cause any problem in practice.
(Why would I try using something I was confident would make me less productive? Because I felt was in a rut, and was trying out disruptive changes to my workflow. For example the same day, I switched all my calculator tools to RPN. That's not gone badly either.)
(Why would I try using something I was confident would make me less productive? Because I felt was in a rut, and was trying out disruptive changes to my workflow. For example the same day, I switched all my calculator tools to RPN. That's not gone badly either.)
What now causes a problem for me is working on code without ligatures, that's after only a few months of working with them. I also find Fira Code extremely legible as a dyslexic.
>In fact , our visual system has been trained to instantly recognize a!=b in code
Citation needed.
>confusing = with == will lead to lots of errors etc
This is more about NOT confusing == with = -- and several similar characters than the opposite.
Sounds like you have thought of all kinds of arguments to not actually try it and see whether it's any good in practice -- which is how anything should be judged (unless it involves some huge risk). In other words, a priori rationalization.
>If I get used to this font on my machine, i will have a real problem when parsing code in another terminal/editor.
The above argument advices to forever stick to a lowest common denominator coding toolset, lest one has to ever try to work in other environment.
How about maximizing the performance one gets from their own, and most frequently used, environment, instead of "what-ifing" for foreign systems they might have to use 1/10 of the time or less?
Plus, having a tool at our disposal and getting used to it doesn't mean we can't also do without if needed. Especially if it's a small improvement, like the one proposed by the ligatures here, and not some totally different notion that we can't ever go back to how it was without it...
Citation needed.
>confusing = with == will lead to lots of errors etc
This is more about NOT confusing == with = -- and several similar characters than the opposite.
Sounds like you have thought of all kinds of arguments to not actually try it and see whether it's any good in practice -- which is how anything should be judged (unless it involves some huge risk). In other words, a priori rationalization.
>If I get used to this font on my machine, i will have a real problem when parsing code in another terminal/editor.
The above argument advices to forever stick to a lowest common denominator coding toolset, lest one has to ever try to work in other environment.
How about maximizing the performance one gets from their own, and most frequently used, environment, instead of "what-ifing" for foreign systems they might have to use 1/10 of the time or less?
Plus, having a tool at our disposal and getting used to it doesn't mean we can't also do without if needed. Especially if it's a small improvement, like the one proposed by the ligatures here, and not some totally different notion that we can't ever go back to how it was without it...
The thing is, in general, if you are promoting some change, you need to show that it brings benefits.
So the onus is on promoter - to show benefits.
The onus is _not_ on everyone in the world to show whether the idea works or not.
So the onus is on promoter - to show benefits.
The onus is _not_ on everyone in the world to show whether the idea works or not.
Promoters (or, to be more precise, creator) might really not care to prove anything. And I wouldn't trust their arguments either.
If the change is harmless enough to try, one can just see for themselves.
And for things that work for some people but not for others, some general proof doesn't make sense anyway.
Nobody came and proved me how vim can work better for my editing than Emacs. I tried both myself and found out. Several for tons of other things that I adopted or not.
If the change is harmless enough to try, one can just see for themselves.
And for things that work for some people but not for others, some general proof doesn't make sense anyway.
Nobody came and proved me how vim can work better for my editing than Emacs. I tried both myself and found out. Several for tons of other things that I adopted or not.
I admit that the >= and <= looked quite more readable, but i don't see how = and == wouldnt get more easily confused. Also i do spend a lot more than 1/10th of time in different environments.
I don't understand why ppl upvoted my comment. You should try it first and report back.
Also, http://www.pallier.org/papers/Chang.Reading.2015.pdf
I don't understand why ppl upvoted my comment. You should try it first and report back.
Also, http://www.pallier.org/papers/Chang.Reading.2015.pdf
If you give it a try, you may find your brain more adaptable than you think. Mine seems able to parse the same code in my editor (with this font) and in web-based repository viewers (with some other monospaced font) without problem.
Anybody else despise the look of many of these ligatures? One of the worst is the design of the ampersand.
I'm assuming this doesn't take care of context?
For eg:
String s = "A != B";
would render the ligature inside the string?
For eg:
String s = "A != B";
would render the ligature inside the string?
I don't think Fira Code recognizes context. It just parses the characters and renders multi-character glyphs into a single one.
So the inequality inside the string will be afforded the same treatment as an inequality anywhere else in the code.
So the inequality inside the string will be afforded the same treatment as an inequality anywhere else in the code.
it doesn't unless you use a different font for strings. But it seems that would be fine in your example
Ligatures aside, I tried this font in IntelliJ and found the letter spacing to be far too wide.
I quite like the "contextual vertical align". Everything else I find either visually unappealing (eg. the ampersand, as already mentioned) or outright confusing (the equality glyphs in particular).
If you want to use Fira Code in Visual Studio Code, use this in settings.json:
"editor.fontFamily": "Fira Code"
"editor.fontLigatures": trueIt looks pretty but totally does not work for my brain.
It doesn't took us long to decode == or >= operator naturally.
It would be cool if it turned .\ into a greek lambda as well.
wow. someone worked really hard to solve a problem that doesn't exit
I'm not sure that I want my equality signs to be ligature'd together. I feel like using this might mess up alignment in a monospace font.
Usually the ligatures in these fonts are designed to fill the same width as the original combination of separate characters in monospace use. That does preserve the alignment of code to either side of the ligature, but unfortunately at the expense of either making some of the symbols much wider than normal or adding a lot of extra whitespace to pad them out. For programming, the ligatured version seems quite risky if it means the only visual difference between single and double equals or underscore characters is a bit of width and/or horizontal padding.
What I would like to try is a language where = actually meant equality rather than assignment and a few other common operators like !=, ===, !==, <= and >= all appeared using the standard mathematical symbols at the same width as =, < and >. That way, some related cases would naturally line up and most lines with comparisons would become shorter and less cluttered, which seems like a modest incremental improvement in language syntax. Alas, you’d also need software and keyboard input to support that in a way that made those alternative symbols effortless and entirely predicable to use. This feels like a job for a standardised programmer’s keyboard and AltGr, except that you’d need a standardised programmer’s keyboard for each of the different languages that already uses AltGr for other purposes, and at some point it all becomes more trouble than it’s worth.
What I would like to try is a language where = actually meant equality rather than assignment and a few other common operators like !=, ===, !==, <= and >= all appeared using the standard mathematical symbols at the same width as =, < and >. That way, some related cases would naturally line up and most lines with comparisons would become shorter and less cluttered, which seems like a modest incremental improvement in language syntax. Alas, you’d also need software and keyboard input to support that in a way that made those alternative symbols effortless and entirely predicable to use. This feels like a job for a standardised programmer’s keyboard and AltGr, except that you’d need a standardised programmer’s keyboard for each of the different languages that already uses AltGr for other purposes, and at some point it all becomes more trouble than it’s worth.
At least with using ligatures in IntelliJ IDEA, I find that it makes zero difference as to how the monospaced font is aligned.
Very interesting but I hate this for programming, feels wrong when I'm so used to 1 char == 1 fixed with char. I'm sure I could adapt but for some reason it makes my sense of programming taste not happy. What do others think?
It takes some time to get used to and it's annoying to place the cursor correctly to backspace from != to =.
But it looks sexy and confuses the co-workers XD. Joke aside: I actually like making my PC mine and like the look. I don't think it makes me a faster or better programmer in any way...
But it looks sexy and confuses the co-workers XD. Joke aside: I actually like making my PC mine and like the look. I don't think it makes me a faster or better programmer in any way...
Does this strike anyone else as a gimmick? I'm not sure if parsing symbols is really the limiting factor when I try to understand code.
It is no more a gimmick than the standard mode of rendering. However, it does have an adaptation cost.
A comparable example: I went to the trouble of creating a custom key binding for my old laptop that was a standard US layout, except the numeral row was symbols by default (i.e. Press shift to get numbers), because I only used the numpad for numerals. It worked, but relearning was too painful. If I'd learned it that way from the beginning, it would just be normal.
Yes, it's a gimmick. The question whether it's a useful gimmick is an open one. I'd try it out probably, but probably it would end more confusing than helpful after the novelty wears off.
Doesn't have to be the limiting factor to still be an improvement.
I like the idea but the xml tags </ /> look horrible
While these are cool for complete code snippets, they are distracting during coding. For example, the colon contextual alignment flickers as the context is typed, and that distraction alone outweighs any convenience of a merged symbol.
Lua's not equals `~=` gets ligatured to ≃ instead of ≠. Expected but unfortunate.
Most of Lua was nicely obvious and intuitive, but ~= was always one of the weird sharp edges. I could maybe see the argument for using it if ~ was already a negation operator of some sort (like C's bitwise complement) but Lua doesn't even have that. Instead you get something needlessly non-intuitive and subject to misinterpretation by anyone who isn't a Lua native, as ~= as an ASCIIfication of ≃ is pretty common.
I've been using this in IntelliJ IDEA for about a month.
The experience reminds me of adopting an anti-aliased font for coding, for the first time (many years ago). First half an hour of unfamiliarity and skepticism, then it just seems very natural, and reading code feels smoother and easier and more productive.
I absolutely cannot stand code in non-monospaced fonts, so I don't think there are any alignment problems. (The look of Fira Mono / Fira Code isn't really to my taste though, so I'm hoping for the same things to start popping up in other fonts too!)
The experience reminds me of adopting an anti-aliased font for coding, for the first time (many years ago). First half an hour of unfamiliarity and skepticism, then it just seems very natural, and reading code feels smoother and easier and more productive.
I absolutely cannot stand code in non-monospaced fonts, so I don't think there are any alignment problems. (The look of Fira Mono / Fira Code isn't really to my taste though, so I'm hoping for the same things to start popping up in other fonts too!)
There are lots of others - my personal favourite is Monoid. http://larsenwork.com/monoid/
So far this font has worked really well for me for Haskell, C++, Python. I find it very readable
This isn't really what I thought it would be. I don't want != changed to ≠ because ≠ isn't a C++ operator.
What I thought would be true ligatures ("tying together") of "&&" which could still be recognized as a pair of ampersands, but by being conjoined cleanly appear to be symbol in their own right, which is what "&&" is.
However part of the value of ligatures is symbol tightness, which is defeated by the need for a monospaced font.
Still, I'm glad people are experimenting!
What I thought would be true ligatures ("tying together") of "&&" which could still be recognized as a pair of ampersands, but by being conjoined cleanly appear to be symbol in their own right, which is what "&&" is.
However part of the value of ligatures is symbol tightness, which is defeated by the need for a monospaced font.
Still, I'm glad people are experimenting!
>I don't want != changed to ≠ because ≠ isn't a C++ operator.
A colored != is not a C++ operator either. It's about getting our minds to differentiate symbols faster, not about whether they exist as is in some language's syntax.
A colored != is not a C++ operator either. It's about getting our minds to differentiate symbols faster, not about whether they exist as is in some language's syntax.
I love typography and ligatures, but this introduces a layer of abstraction in your source code. I know it sounds depressingly uninternational, but I don't even like to use Unicode characters in my text files. I prefer a one-to-one mapping of keyboard button to on-screen character. Although I suppose even capital letters break that rule, so maybe one day I'll give it a whirl.
If I get used to this font on my machine, i will have a real problem when parsing code in another terminal/editor. Unless all the programmers in the world decide to switch to it simultaneously, this introduces a mental translation that's counterproductive.