Yes, but if it to look like you intended then the size of the tabstop is critical. I mean if you assume the tabstop is 2 characters and I view it with tabstops of 8 then the appearance is very different. Also you are assuming that you start the indentation with tabs and then extend it with spaces. Mixing tabs and spaces does not work.
I still think the easiest and most consistent is to always use spaces in the file and have a smart tab key which knows about indentation and goes to the right place.
To me the ‘intended size’ can be anything, that is why I only use tabs for indention, sometimes I may view them at 8-tabstop, sometimes 2, even I change on occasion, with using that pattern it just ‘always works’ without needing to reformat the code. ^.^
And yes, start indentation with tabs, then do alignment with spaces. Tabs will only ever exist at the very start of a line with a count of 0-to-N, never elsewhere.
That is horror on formatting though. What about a project that uses 8-spaces as their tabs and you want to print it (manager stupidity or something), or what if a project uses 2 but you’ve not had coffee yet and want to view it with more? Smart tab keys are fine for putting ‘in’ auto-spaced indentation, but it does nothing for viewing existing files. Using a mix of tabs and spaces is forward thinking as it allows you to both put in with ease and view an arbitrary-period-of-time-in-the-future with ease on whatever display you may have then.
// 4-tabstop
int f(int x, // line 1
......int y) { // line 2
--->return g(x, // line 3
--->.........y); // line 4
}
When discussing functional programming the “code that is easier to reason about” argument invariably surfaces - on the flip side when writing code I don’t want to have to waste any brain cycles to determine where indentation stops and alignment starts (e.g. suppress the urge to use the tab key on line 2 and to use more spaces than my tab width to ensure proper alignment on line 4). I guess this type of formatting could be left to the editor - if one can be bothered to getting it properly configured for such a task.
There is more to coding styles than just indentation - maybe I find it easier to read code like this
int f(
int x,
int y
) {
return g(
x,
y
);
}
Ultimately we can’t escape the author’s/organization’s coding style short of reformatting the code either automatically or manually - and often there are far greater obstacles to code readability than just tab-size.
According to the article on github spaces win over tabs - not that majorities are always right (they are just an indication of what’s popular) but it certainly lines up with the notion that with “just spaces” there is less to keep track of.
Interestingly github lets you control your displayed tab-size with a ts query string parameter, e.g.:
For Elixir work I often still use Atom, which does not have a good plugin for that, but for Elixir I keep to 2-spaces for indentation regardless in all code as that is part of the style of the language.
Computer programmers who use spaces as part of their coding earn $15,370 (£12,000) more per year than those who use tabs, a survey of developers has revealed.
The survey found the salary difference stretched across different languages, countries and experience levels.
The debate over whether it is better to use spaces or tabs to indent code has raged among programmers for years.
Indents act like paragraph markers and help define how programs work.
The result was “surprising,” said David Robinson, data scientist at Stack Overflow which carried out the survey of 12,400 developers.
I tend to use spaces, because I have much more control over how it will look like for others.
I do use tabs though where I have to, be it because its considered more idiomatic (go), because they are needed to actually work (GNU Makefile) or just because they are corporate policy (complicated[*]).
[*] In my current job we prefer 2 spaces, but since we also use go and GNU Make, we use a tabwidth of 4 there and also we use a CI tool which ensures whitespace consistency.
Did you read the actual post? They controlled for that and saw the same effect. (Or are you referring to how certain languages were excluded from the chart? Did anyone run the numbers themselves from the survey data without a filter?)
Where are you getting that? I don’t think the SO survey asks about typography. (I may be missing a joke. )
