I use BEM – A good clear naming convention

It’s not the most exciting skill in the world, but I like to think I’m pretty good at naming things. That’s because it’s so damned important to be clear and concise, or you risk confusion.

That’s why I use BEM.

Bem is great, although can get a bit wordy with massive long class names. But I try not to let it. It’s great because you can declare a reusable component, for example:

search {}

search__input {}

search__btn {}

search__label {}

And then use a standard BEM modifier (“–“) to make alternate versions, just by adding a new class:

search–vertical {}

…inheriting most of the styles but adding some overrides.

The best thing about BEM though is the specificness, and you’re absolutely certain the class hasn’t been used elsewhere. For example if I found a class like “full-width” I wouldn’t dare touch the css for it, in case it was inherited by something else tucked away on the website. But I know that “search–full-width” It’s pretty clear what that class is for and where it’ll be used. It’s just clear naming really.

My CSS layout – group by component, not by screen-size

I read an interview with Eric A Meyer today, who is supposedly the “Godfather of CSS wisdom”, and one thing that stuck out for me is how he structures his CSS files. He says:

I usually organize my CSS by section of the layout, so the header styles are all together at the top of the CSS, the page-content styles are in the middle, and the footer styles come at the end. Then I do my media-query blocks, with the same internal organization.

No, no, no Eric! The first bits seem fine, but don’t group your CSS by screen sizes into different media queries! I have to say I’m a little bit confused why anyone would do it that way, splitting your entire stylesheet into sections wrapped in different media queries. It’s like building the website multiple times over.

I tried to explain this in an older article (see it here), but wasn’t too clear so I’ll have a go at explaining the reasons why I think you should avoid grouping by screen size, and what I do instead, which I actually think is much easier (although obviously do whatever suits you and your project!).

1) Working on the file – and tracking down the distant sections you need

Okay so I’m assuming you’ve got a mobile-first website (and if you’re not using mobile first, you should be!). And if you’re doing it Eric’s way you’ve got at least 2 separate chunks of CSS, styles for mobile (including default styles), and the styles for desktop separately. Between them you might also have media queries for in-between, tablets or small browser windows on desktop. So if you’re using Eric’s method, to change any component on the page, say… the footer, you’d have to skip between distant sections of the same document to add styles.

2) What if you want to target a different screen size?

Okay so you have the very specific styles for mobile, tablet, small and large desktop screens. But what about this, a new device comes out which your client/boss tells you is a big deal. The website HAS to look awesome at this particular size and aspect ratio. Not just acceptable – awesome.

So Eric’s method, he has to create an entire new media query section, slipping it in between the others. What does this do to your mobile first approach? Well you’ll have to leech some of the styles from the larger media query perhaps, and maybe add a few more very specific ones for that size range. Seems like your nice flow of mobile first to larger has just got a bit more bumpy.

My flow – 1) Generic to specific

The method I use was taught my CSS Wizardry’s Harry Roberts in a workshop in London a few years ago. I’ve changed it slightly to suit me, but I think the flow is pretty solid.

He starts with a triangle, because that’s a good way to imagine the best way to write CSS, from the most generic to the most specific, ending with some “trumps” which is essentially allowing for some hacks right at the end.

harry roberts css flow

His triangle needs a bit of explaining – Settings & tools are where you’d define variables and include CSS mixins, maybe your frameworks. Objects and components are things which are reusable throughout the website. Harry explains the rest here, go read that because it’s excellent, and it allowed me to come up with what works for me…

My flow – 2) Files I include

  • Framework
    • For me this is Bootstrap
  • Variables
  • Mixins
  • Animations
  • Global styles (fonts etc)
  • Components
    • Header
    • Footer
    • Menus
    • etc
  • Specific layouts
    • Home
    • About
    • etc

This order allows you to build reusable components, then if you need a specific alternative for a certain page, you can make a slightly more specific version without having to create a brand new version.

An example of this is all buttons (I like to use the class .btn) might be the same size colour and padding, except on the login page where you can make them all a bit bigger easily using a bit more specificity.

.login-layout .btn {}

Including separate SCSS files

Each of the above is a separate SCSS file, included into the styles.scss, which is where I’ll have global styles for typeface, font sizes, all the basic stuff for the entire site. This works for every project I’ve done recently.

I compile all of these into a single css file, including the entirity of Bootstrap which I do so I can extend some of its classes and use its mixins. It’s all included in the order above, with the most generic items able to be extended further down the line.

One cool thing about SCSS is that if you prefix the filename with an underscore, it doesn’t try and compile the file separately, it know it’s an include file only.

But what about the media queries?

One thing I didn’t mention in the list above is the media queries. That’s because I don’t have a specific place for those. They’re intermingled around the entire thing. If you’re not used to that, you might think it madness. But honestly it feels a lot more organised than the alternative. If you want to work on the _header.scss, all the header styles are right there in one place, and you don’t have to go chasing them down.

I use BEM too, but I’ll explain that in another article!

Mystery meat & hidden salad

Mystery meat – A term given for an option that you don’t know what it’ll do until you choose/select/click it. The metaphor as I underswtand it: You’re choosing tinned meat, but don’t know what’s inside until you buy it and open it.

I especially hate this used in navigation, where in recent years it’s become common practice to replace words with icons, to easily translate the interface into different languages, and fit more onto the screen for mobile.

Hidden salad – I went to a resetaurant recently and asked for a side salad with my burger instead of fries. The burger wasn’t supposed to come with a salad, so I expected to pay extra. As it turns out, the salad WAS a legitimate option, but a hidden one.

“Hidden salad” is what I call options that are valid but completely hidden from the user. Similar to mystery meat but instead of being presented with a cryptic option, you’re not given any option at all until you discover it. This is used especially in interfaces where you have to type, like command-line tools (which I especially hate). Or a more fun use might be Google’s hidden easter eggs.

WordPress have ditched the title tag in links

WordPress have previously always asked for title text when adding a link, but from version 4.2 they’ve abandoned this.

LinkmodalChange…Which is quite a good thing imo.

I wrote an article about how bad the title tag is for accessibility, because some web developers assume this is read aloud by screenreaders, bit it isn’t. I wrote an article about this here: I thought title tag improved accessibility, I was wrong.

I learned this after pretending to be blind for a week, and learning how to use a screenreader. I learned loads, and recommend every web developer to do the same!

Image above is from this article which is where I was first alerted about WordPress’ change.