Post pobrano z: Thank You (2017 Edition)

As 2017 comes to a close, as we do each year, let’s take a numbers-based glance back at CSS-Tricks. And more importantly, tip our collective hat to all y’all that come here and read the site.

We really do think of the site as somewhere you come and read. While a lot of people’s experience with CSS-Tricks is a place that shows up in search results to find the answer to some web design/dev question (and that’s awesome), another way to experience the site is to read it like a magazine. We publish an article (or a couple) nearly every day, from a variety of authors, with the hope that it’s interesting and exposes us all to new ideas.

According to Google Analytics, which we’ve had installed and reported from anonymously since day 1 around here, we had 75 million pageviews this year. Very little of CSS-Tricks makes use of any kind of „single page app” type tech, so that’s pretty accurate to people visiting and clicking around. It’s down from 77 million last year. I’d like to think that’s because of ad blockers, which often block Google Analytics, are up in usage for the visitors of CSS-Tricks, but it’s just as likely we’re just down a smidge this year. Sessions are also down a smidge at 54 million but Users steady at 21 million.

We were on a publishing role though! We published 595 posts, blowing away last year with only 442, the previous record. We also published 50 pages (i.e. snippets/videos/almanac entries) beating 43 last year. Certainly, we’re in favor of quality over quantity, but I think this is a healthy publishing pace when our goal is to be read, in a sense, like a magazine. That has been more clear to me this year. We produce content with complementary goals and one of those goals is that of a magazine. We hope the site is worth reading day after day and week after week. The other is that the content lasts and is referenceable for many years to come. Hopefully thinking of things that way can be a guiding light, balancing news and reference content, while avoiding stuff that is neither.

I always wished there was an easy way to figure out what the most popular articles published that year were, but I still don’t have a great way to do that. The top 10 is dominated by our big guides, things like our Guides to Grid, Flexbox, SVG, and centering.

Those top 10 make up about 15% of total traffic, which is a massive slice, but that leaves 85% of traffic as part of the „long tail”. That’s a satisfying thought when you’re in it for the long haul as we are. Not every article is a huge top 10 smash hit, but does contribute to the long tail which is a much bigger slice collectively anyway.

For the last bunch of months, we’ve been using Algolia for search. My plan only has 7 days of analytics retention, which isn’t enough data to expose major trends. In looking at a week of data though, you can see some surprising top terms like React, SVG, grid, flexbox, font, border… Another thing that’s clear is that on-site search is quite important. Algolia reports ~15,000 queries a day. I don’t think that represents „user typed in a thing and submitted it” because the search happens in real-time, so each character typed can result in a query. Still, likely hundreds or low-thousands of searches a day.

I’m going to convert us back to using Google search. I think Algolia is a fantastic product, I just don’t have the developer time right now to give it the work it needs.

The location of y’all seems to be spreading out internationally little by little. The United States is down to 22% of traffic from 23% and India back to down to 11% from 12% (meaning more traffic came from elsewhere). Those are the top 2, then it goes UK, Germany, Canada, France, Netherlands, Poland, Russia, Brazil. What would be really interesting is to figure out visitors per-capita. For example, Singapore has a population of 5.6 million and had 111,319 unique users, so you could say about 2% of Singaporeans visited CSS-Tricks last year. Lols probably not, but hey that’s what the back-of-the-napkin math shows. Whereas with the 4.6 million unique visitors from the US compared to the 323 million population means only 1.5% has visited.

We gained about 10,000 newsletter subscribers this year for a total of 31,376. That’s about a third of the entire list size. I love our newsletter. I think it has a ton of potential and is always worth reading. To be perfectly honest I’d like to see our newsletter subscriber numbers be higher than our Twitter followers, but that’ll be a tall hill to climb.

Search is the origin of 86.6% of the traffic we get. Direct visits and referral links result in another 5% each. Social media just 2.5%. Whenever I look at that I’m reminded of the disproportionate amount of energy spent on it. Still, it aligns with our goal of being a publication people regularly read and being a source of news, so it feels worth it.

Speaking of social media, we rose 44,000 follows on Twitter last year, again an astonishing number, but it’s down year-over-year for the last several years. 71,900 likes on Facebook, only rising about 3,000, which isn’t bad considering we do hardly anything at all on Facebook. Growth is much more explosive on YouTube. We’re at 40,123 subscribers there from 32,174 last year, despite only posting a paultry 6 videos.

This is a classic loop in my head: maybe we should have a dedicated social media person! But even part-time employees are expensive… is it worth it? How much more potential is there? Could they add so much value they pay for themselves? What else could they do? And then the questions swirl around in my head so quickly the idea fizzles out. I guess I’ll just say if that kind of thing interests you, reach out!

For once, mobile traffic is actually up. 6.2% this year from below 5% last year. Industry-wide, that’s rock bottom. CSS-Tricks is just weird that way. A lot of developers b searching stuff at work, unsurprisingly. Less than 1% is tablets. 30% of the mobile traffic is iPhones.

Y’all left about 5,040 comments on the site this year, which is a smidge down year over year from the past few years, but it actually feels significantly up to me, since we’ve been much more stringent about what comments we publish this year. I do the vast majority of comment moderation and I trash a lot more than I used to. Anything that off-topic, rude, or unhelpful doesn’t see the light of day. I hope that doesn’t scare you off from commenting. In fact, I hope it encourages it. Anything on-topic and helpful will absolutely be published and won’t be lost in a sea of junk.

We’ve had 20,238 people use our contact form. Don’t worry, I love email.

Goal Review

Double newsletter subscribers. We didn’t double but we grew by a third, which is pretty strong. Especially since we did little to nothing to promote it. We probably need to do a better job of promoting it and somehow incentivizing signups, especially since it’s such a good way to reach people.

More pairing videos. Pretty hard fail here. The main difficulty is scheduling multiple people together, combined with the pressure of producing something worth watching. It’s one thing for an audio podcast like ShopTalk where we can schedule someone and just chit-chat about tech. It’s another thing to ask someone to pair with you and essentially do live coding. It’s still a good idea, it just needs a more serious time commitment and leader. And honestly, probably a sponsor so that it can be worth everyone’s time.

Maintain a mostly-daily publishing schedule. Check and check! This is the first year we’ve actually kept an honest to god schedule, structured after daily posting. We’ll be moving forward with that for sure.

Assemble content in more useful ways. We got a good start on this with Guides. We haven’t done a ton with them yet, but we have given ourselves a way to build these without too much effort, and I think the potential in them is fantastic.

New Goals

Publish something in a new format. We have a lot of publishing power around here with lots of writers and a solid platform. Let’s use it to publish a book or something book-like.

More editorial vision. Admittedly, what we published each day is a bit random. That’s not a terrible thing since the site has a niche anyway, but I’d call it a good goal to exert some editorial guidance to what we publish and when. Meaning commissioning and shepherding work that is a good fit for this site and publishing it with timing that makes some sort of sense.

Interesting sponsorship partners. The most rewarding and powerful partnerships between sponsors and publications are ones of mutual respect and admiration. There are loads of businesses out there I think are doing a terrific job of building what the build, and I’d like to forge relationships with them to promote what they do. And do that promotion in a way that we are uniquely able to do. Get in touch if you think you’re a part of that company.

Create another very popular page. It’s just a matter of time and topic. I’d like to find at least one web development topic that could really use a strong reference page and spend a good amount of time making what we consider a perfect fit for that, with the goal of it really resonating with developers.

Most importantly

Thank you, again, for being a part of this community.

See the Pen Text Animation with CSS – thank you screen! by Nooray Yemon (@yemon) on CodePen.

Thank You (2017 Edition) is a post from CSS-Tricks

Post pobrano z: What You Build

I tweeted this yesterday and it seemed to resonate with some folks:

Just a little reminder that it’s about 100 times more important what you build than how you build it.

— Chris Coyier (@chriscoyier) December 10, 2017

What I was feeling when I wrote that was a little tired of endless discussions on tech minutia and yearning for more focus on what we are building and discussion about why.

If you’re a reader of this site, you and I live in the same bubble. It’s a nice bubble. It’s full of smart people who like to chat about web design and development. I live it and love it.

It’s easy to get into a heated discussion about frameworks, what type of class names make the most sense, which optimization techniques are most important, or what part of your code base is should be responsible for styling. Those are great discussions that guide our industry.

But what is more important? The naming convention you chose or if your user can actually book a flight? Which state store library you picked or if you actually had the scarf your user was looking for? Which command line tool pulled your dependencies or whether someone was able to find and read the instructions to send in their court appeal?

I was trying to encourage people to build and think about what they are building rather than get too discouraged about the how. You’re building things for people and that’s such a big responsibility. One that outweighs technical choices, as important as they seem.

I enjoyed the pushback I got on it though.

Most of it centered around the fact that if you make poor tech choices, that limits the quality of what you build and slows your ability to change and adapt to changing user needs. Fair enough.

Good tech just might lead to directly better features and UX for your users. Fair enough. Good tech might be a differentiator between you and your competition. Fair enough.

My favorite was calling out the story of the three little pigs. If you aren’t familiar, there is a Big Bad Wolf that is trying to eat the piggies. Each of them built a house to protect themselves. I imagine you can guess which of the pigs did better: the one who built their house out of hay, or the pig who built their house out of bricks.

Fair enough.

Drew McLellan gets into this in All That Glisters, but focuses on the old vs new tech issue:

There are so many new tools, frameworks, techniques, styles and libraries to learn. You know what? You don’t have to use them. You’re not a bad developer if you use Grunt even though others have switched to Gulp or Brunch or Webpack or Banana Sandwich. It’s probably misguided to spend lots of project time messing around with build tool fashions when your so last year build tool is already doing what you need.

And this gem:

Software, much like people, is born with a whole lot of potential and not much utility. Newborns — both digital and meaty — are exciting and cute but they also lead to sleepless nights and pools of vomit.

He goes on to say that what you are building might help dictate your tech choices. Ah yes, the what. Not only is what your things does litearlly the only thing people care about, it also helps guide tech choices.

What You Build is a post from CSS-Tricks

Post pobrano z: You are what you document

There are so many little gems in this piece by Yevgeniy Brikman all about documentation. He digs into a lot more than simply documenting code though and focuses on how we can document every phase of our work, from design to process and beyond.

Here’s my favorite lines that made me sit back and shout “Wahoo!”:

When a developer uses your code, they are really learning a new language, so choose the words in it wisely.

…programs must be written for people to read, and only incidentally for machines to execute.

I like how Yevgeniy suggests that there are kinda two different mindsets that we have to get into when writing code: one for making the dang thing work in the first place, and another for explaining how and why we did a specific way. There’s context-switching that takes place between those different stages of work.

Anyway, when seen in this light, documentation could be much more than a nice-to-have. Instead, it might just be 50% of the work.

Direct Link to Article — Permalink

You are what you document is a post from CSS-Tricks

Post pobrano z: Additive Animation with the Web Animations API

These features have not landed in any stable browsers at the time of writing. However, everything discussed is already in Firefox Nightly by default and key parts are in Chrome Canary (with the Experimental Web Platform Features flag enabled), so I recommend using one of those browsers (when reading this article) to see as many of the features in action as possible.

Regardless your preferred method of animation on the web, there will be times that you need to animate the same property in separate animations. Maybe you have a hover effect that scales an image and a click event that triggers a translate — both affecting the transform. By default, those animations do not know anything about the other, and only one will visually be applied (since they are affecting the same CSS property and the other value will be overridden).

element.animate({
  transform: ['translateY(0)', 'translateY(10px)']
}, 1000);

/* This will completely override the previous animation */
element.animate({
  transform: ['scale(1)', 'scale(1.15)']
}, 1500);

The second animation in this Web Animations API example is the only one that would be visually rendered in this example as both animations play at the same time and it was the last one defined.

Sometimes we even have grander ideas where we want a foundational animation and then based on some user interaction change in state we smoothly modify the animation a bit midway without affecting its existing duration, keyframes, or easing. CSS Animations and the current Web Animations API in stable browsers today cannot do this out of the box.

A New Option

The Web Animations specification introduces the composite property (and the related iterationComposite). The default composite is 'replace' and has the behavior we have had for years now where an actively animating property’s value simply replaces any previously set value — either from a rule set or another animation.

The 'add' value is where things change from the previous norms.

element.animate({
  transform: ['scale(1)', 'scale(1.5)']
}, {
  duration: 1000,
  fill: 'both'
});
element.animate({
  transform: ['rotate(0deg)', 'rotate(180deg)']
}, {
  duration: 1500,
  fill: 'both',
  composite: 'add'
});

Now both animations will be seen as the browser on the fly figures out the appropriate transformation at a given point in the element’s timeline accounting for both transformations. In our examples, the easing is 'linear' by default and the animations start at the same time, so we can break out what the effective transform is at any given point. Such as:

• 0ms: scale(1) rotate(0deg)
• 500ms: scale(1.25) rotate(60deg) (halfway through first animation, 1/3 through second)
• 1000ms: scale(1.5) rotate(120deg) (end of first, 2/3 through second)
• 1500ms: scale(1.5) rotate(180deg) (end of second)

See the Pen Animation Composite by Dan Wilson (@danwilson) on CodePen.

So Let’s Get Creative

An individual animation does not just consist of a start state and end state — it can have its own easing, iteration count, duration, and more keyframes in the middle. While an element is mid animation you can throw an additional transformation on it with its own timing options.

See the Pen Add more transform animations by Dan Wilson (@danwilson) on CodePen.

This example lets you apply multiple animations on the same element, all affecting the transform property. To keep from going all out in this example, we limit each animation to a single transformation function at a time (such as only a scale), starting at a default value (such as scale(1) or translateX(0)), and ending at a reasonable random value on that same transformation function, repeated infinitely. The next animation will affect another single function with its own randomized duration and easing.

element.animate(getTransform(), //e.g. { transform: ['rotate(0deg), 'rotate(45deg)'] }
{
  duration: getDuration(), //between 1000 and 6000ms
  iterations: Infinity,
  composite: 'add',
  easing: getEasing() //one of two options
});

When each animation starts, the browser will effectively find where it is in its previously applied animations and start a new rotation animation with the specified timing options. Even if there is already a rotation going in the opposite direction, the browser will do the math to figure out how much a rotation needs to happen.
Since each animation has its own timing options, you are unlikely to see the exact same motion repeated in this example once you have added a few. This gives the animation a fresh feel as you watch it.

Since each animation in our example starts at the default value (0 for translations and 1 for scaling) we get a smooth start. If we instead had keyframes such as { transform: ['scale(.5)', 'scale(.8)'] } we would get a jump because the didn’t have this scale before and all of a sudden starts its animation at half scale.

How are values added?

Transformation values follow the syntax of in the spec, and if you add a transformation you are appending to a list.

For transform animations A, B, and C the resulting computed transform value will be [current value in A] [current value in B] [current value in C]. For example, assume the following three animations:

element.animate({
  transform: ['translateX(0)', 'translateX(10px)']
}, 1000);

element.animate({
  transform: ['translateY(0)', 'translateY(-20px)']
}, { 
  duration:1000,
  composite: 'add'
});

element.animate({
  transform: ['translateX(0)', 'translateX(300px)']
}, { 
  duration:1000,
  composite: 'add'
});

Each animation runs for 1 second with a linear easing, so halfway through the animations the resulting transform would have the value translateX(5px) translateY(-10px) translateX(150px). Easings, durations, delays, and more will all affect the value as you go along.

Transforms are not the only thing we can animate, however. Filters (hue-rotate(), blur(), etc) follow a similar pattern where the items are appended to a filter list.

Some properties use a number as a value, such as opacity. Here the numbers will add up to a single sum.

element.animate({
  opacity: [0, .1]
}, 1000);

element.animate({
  opacity: [0, .2]
}, { 
  duration:1000,
  composite: 'add'
});

element.animate({
  opacity: [0, .4]
}, { 
  duration:1000,
  composite: 'add'
});

Since each animation again is 1s in duration with a linear easing, we can calculate the resulting value at any point in that animation.

• 0ms: opacity: 0 (0 + 0 + 0)
• 500ms: opacity: .35 (.05 + .1 + .2)
• 1000ms: opacity: .7 (.1 + .2 + .4)

As such, you won’t be seeing much if you have several animations that include the value 1 as a keyframe. That is a max value for its visual state, so adding up to values beyond that will look the same as if it were just a 1.

See the Pen Add more opacity animations by Dan Wilson (@danwilson) on CodePen.

Similar to opacity and other properties that accept number values, properties that accept lengths, percentages, or colors will also sum to a single result value. With colors, you must remember they also have a max value, too (whether a max of 255 in rgb() or 100% for saturation/lightness in hsl()), so your result could max out to a white. With lengths, you can switch between units (such as px to vmin) as though it is inside a calc().

For more details, the specification outlines the different types of animation and how the result is calculated.

Working with Fill Modes

When you are not doing an infinite animation (whether you are using a composite or not) by default the animation will not keep its end state as the animation ends. The fill property allows us to change that behavior. If you want to have a smooth transition when you add a finite animation, you likely will want a fill mode of either forwards or both to make sure the end state remains.

See the Pen Spiral: Composite Add + Fill Forwards by Dan Wilson (@danwilson) on CodePen.

This example has an animation with a spiral path by specifying a rotation and a translation. There are two buttons that add new one second animations with an additional small translation. Since they specify fill: 'forwards' each additional translation effectively remains part of the transform list. The expanding (or shrinking) spiral adapts smoothly with each translation adjustment because it is an additive animation from translateX(0) to a new amount and remains at that new amount.

Accumulating animations

The new composite option has a third value — 'accumulate'. It is conceptually in line with 'add' except certain types of animations will behave differently. Keeping with our transform, let’s start with a new example using 'add' and then discuss how 'accumulate' is different.

element.animate({
  transform: ['translateX(0)', 'translateX(20px)']
}, {
  duration: 1000,
  composite: 'add'
});
element.animate({
  transform: ['translateX(0)', 'translateX(30px)']
}, {
  duration: 1000,
  composite: 'add'
});
element.animate({
  transform: ['scale(1)', 'scale(.5)']
}, {
  duration: 1000,
  composite: 'add'
});

At the 1 second mark (the end of the animations), the effective value will be:

transform: translateX(20px) translateX(30px) scale(.5)

Which will visually push an element to the right 50px and then scale it down to half width and half height.

If each animation had been using 'accumulate' instead, then the result would be:

transform: translateX(50px) scale(.5)

Which will visually push an element to the right 50px and then scale it down to half width and half height.

No need for a double take, the visual results are in fact the exact same — so how is 'accumulate' any different?

Technically when accumulating a transform animation we are no longer always appending to a list. If a transformation function already exists (such as the translateX() in our example) we will not append the value when we start our second animation. Instead, the inner values (i.e. the length values) will be added and placed in the existing function.

If our visual results are the same, why does the option to accumulate inner values exist?

In the case of transform, order of the list of functions matters. The transformation translateX(20px) translateX(30px) scale(.5) is different than translateX(20px) scale(.5) translateX(30px) because each function affects the coordinate system of the functions that follow it. When you do a scale(.5) in the middle, the latter functions will also happen at the half scale. Therefore with this example the translateX(30px) will visually render as a 15px translation to the right.

See the Pen Visual Reference: Transform Coordinate Systems by Dan Wilson (@danwilson) on CodePen.

Therefore, with accumulation we can have an order that is different then when we always append the values to the list.

Accumulating for Each Iteration

I mentioned before that there is also a new related iterationComposite property. It provides the ability to do some of the behaviors we have already discussed except on a single animation from one iteration to the next.

Unlike composite, this property only has two valid values: 'replace' (the default behavior you already know and love) and 'accumulate'. With 'accumulate' values follow the already discussed accumulation process for lists (as with transform) or are added together for number based properties like opacity.

As a starting example, the visual result for the following two animations would be identical:

intervals.animate([{ 
  transform: `rotate(0deg) translateX(0vmin)`,
  opacity: 0
}, { 
  transform: `rotate(50deg) translateX(2vmin)`,
  opacity: .5
}], {
  duration: 2000,
  iterations: 2,
  fill: 'forwards',
  iterationComposite: 'accumulate'
});

intervals2.animate([{ 
  transform: `rotate(0deg) translateX(0vmin)`,
  opacity: 0
},{ 
  transform: `rotate(100deg) translateX(4vmin)`,
  opacity: 1
}], {
  duration: 4000,
  iterations: 1,
  fill: 'forwards',
  iterationComposite: 'replace' //default value
});

The first animation is only bumping up its opacity by .5, rotating 50 degrees, and moving 2vmin for 2000 milliseconds. It has our new iterationComposite value and is set to run for 2 iterations. Therefore, when the animation ends, it will have run for 2 * 2000ms and reached an opacity of 1 (2 * .5), rotated 100 degrees (2 * 50deg) and translated 4vmin (2 * 2vmin).

See the Pen Spiral with WAAPI iterationComposite by Dan Wilson (@danwilson) on CodePen.

Great! We just used a new property that is supported in only Firefox Nightly to recreate what we can already do with the Web Animations API (or CSS)!
The more interesting aspects of iterationComposite come into play when you combine it with other items in the Web Animations spec that are coming soon (and also already in Firefox Nightly).

Setting New Effect Options

The Web Animations API as it stands in stable browsers today is largely on par with CSS Animations with some added niceties like a playbackRate option and the ability to jump/seek to different points. However, the Animation object is gaining the ability to update the effect and timing options on already running animations.

See the Pen WAAPI iterationComposite & composite by Dan Wilson (@danwilson) on CodePen.

Here we have an element with two animations affecting the transform property and relying on composite: 'add' — one that makes the element move across the screen horizontally and one moving it vertically in a staggered manner. The end state is a little higher on the screen than the start state of this second animation, and with iterationComposite: 'accumulate' it keeps getting higher and higher. After eight iterations the animation finishes and reverses itself for another eight iterations back down to the bottom of the screen where the process begins again.

We can change how far up the screen the animation goes by changing the number of iterations on the fly. These animations are playing indefinitely, but you can change the dropdown to a different iteration count in the middle of the animation. If you are, for example, going from seven iterations to nine and you are seeing the sixth iteration currently, your animation keeps running as though nothing has changed. However, you will see that instead of starting a reverse after that next (seventh) iteration, it will continue for two more. You can also swap in new keyframes, and the animation timing will remain unchanged.

animation.effect.timing.iterations = 4;
animation.effect.setKeyframes([
  { transform: 'scale(1)' },
  { transform: 'scale(1.2)' }
]);

Modifying animations midway may not be something you will use every day, but since it is something new at the browser level we will be learning of its possibilities as the functionality becomes more widely available. Changing iteration counts could be handy for a game when a user get a bonus round and gameplay continues longer than originally intended. Different keyframes can make sense when a user goes from some error state to a success state.

Where do we go from here?

The new composite options and the ability to change timing options and keyframes open new doors for reactive and choreographed animations. There is also an ongoing discussion in the CSS Working Group about adding this functionality to CSS, even beyond the context of animations — affecting the cascade in a new way. We have time before any of this will land in a stable major browser, but it is exciting to see new options coming and even more exciting to be able to experiment with them today.

Additive Animation with the Web Animations API is a post from CSS-Tricks