Forum comments in chronological order

Anonymous
Tue 7-Mar-2023 14:49

I came here through http://www.rkoucha.fr/tech_corner/pty_pdip.html
Both excellent contents, and this page is running since 2008!!!

Congratulations

Anonymous
Fri 17-Mar-2023 16:20

I can't speak as a true programmer/developer, I do sysadmin work in PowerShell mostly. But syntax highlighting is very helpful as a beginner/intermediate coder. It helps ensure that me and PowerShell are both interpreting the code the same way - if it's highlighted, it's recognized by the shell as the part of syntax that I intended it to be. If it's not, I've made a mistake.

I do think there is some validity to the argument against highlighting though, so I'm going to give it a try and see what I prefer. Thanks for posting this, and thanks to everyone who commented!

Anonymous
Sun 19-Mar-2023 13:49

Linus,

Thank you for writing such an informative article. All the best!

-Lane F.

Anonymous
Fri 31-Mar-2023 20:36

I'm sure good arguments can be made against syntax highlighting. But this was not it. Especially setting up straw-mans (like: syntax highlighting is like training wheels, training wheels are bad, therefore syntax highlighting is bad) doesn't help your case.

It's easy to prove that syntax highlighting is not really essential to programming. Just look at blind programmers, or undeniably great programmers like Linus Torvalds who uses MicroEMACS, an editor which not only lacks syntax highlighting but many other features most programmers would consider "essential".

But it's also proven that it can help speed up development (see https://ppig.org/papers/2006-ppig-18th-hakala/ or https://ppig.org/papers/2015-ppig-26th-dimitri/, although one would wish for studies with bigger samples). The data seem to show that the more experienced the programmer, the less the impact of syntax highlighting, but that doesn't automatically make them "training wheels" as long as they still help.

Going through a compile cycle to detect a syntax error, is much worse than detecting and fixing the error on the spot. But I'd argue that syntax errors are not really the case for syntax highlighting (they can be pointed out by the semantic analyzer in a modern IDE). Their advantages are more on the semantic comprehension side. E.g. to come back to your `=`/`==` comparison: I don't know which IDE you use, but at least in my Emacs setup the coloring of the left side variable clearly changes with a switch from `=` to `==`, which helps to clearly distinguish lvalues in assignments from values in predicates (changing the coloring of the operator itself would not only have a barely noticeable effect but also wouldn't help much).

Between being blind and addicted to colors, there is a whole spectrum of people. So we can't make generalizing statements like "Syntax highlighting is good for everyone". But to make the opposite case, you'd need much better arguments and thorough research.