A designer I work with was presenting comps at a recent team meeting. She had done a wonderful job piecing together the concept for a design system, from components to patterns and everything in between that would make any front-end developer happy.
But there was a teeny tiny detail in her work that caught my eye: the hover state for links was a squiggle.
Default link (top) and hover effect (bottom)
Huh! Not only had I not seen that before, the idea had never even crossed my mind. Turns out there are plenty of instances of it on live sites, one being The Outline. That was the one that was implementation that inspired the design.
Cool, I figured. We can do something like a linear background gradient or even a background image. But! That wasn’t the end of the design. Turns out it’s animated as well. Again, from The Outline:
Whoa! That’s pretty wild. I wasn’t sure how to approach that, honestly, because animating any of my initial ideas would be difficult, especially for cross-browser support.
So, how did The Outline do it? Turns out, it’s SVG. We can make a squiggly path and animate it pretty easily:
But that’s kinda crappy because we can’t really animate that. We need better values for that. However, we can inline CSS directly on the SVG in the background-image property. We can’t simply copy and paste the SVG code into the property, but we can with some proper encoding:
And, since SVG can contain its own styles within the markup, the animation can be tossed right there in the background-image property, the same way we would do it with CSS in an HTML document head or inline CSS in HTML.
I have no idea if an animated squiggle makes for a good user experience and, frankly, that’s not the point of this post. The point is that The Outline had a fun idea with a slick CSS implementation.
That got me thinking about other non-standard (perhaps even unconventional) hover styling we can do with links. Again, tossing aside usability and have a grand ol’ time with CSS…
The Border to Background Effect
Maybe that same bottom border on the default link can grow and become the full background of the link on hover:
Security researcher Sabri posted a bit of code that will „force restart any iOS device.” It’s interesting to see HTML & CSS have this kind of dangerous power. It’s essentially a ton of <div>s scaled to be pretty huge and then set over a repeating JPG image with each <div> blurring the background via backdrop-filter. It must cause such extreme and unhandled memory usage that it wreaks havoc on the browser as well as the entire operating system.
I was trying to test it out myself and be really careful not to execute it… but of course I did, and it crashed my Chrome 68 on a MacBook Pro. Not the whole operating system, but I had to force quit the browser. Then again, I suppose even while(true) {} can do that!
If you put a link on cats.html that links to /dogs.html (an „absolute” path), it’s going to 404 — there is no dogs.html at the base/root level of this site! The / at the beginning of the path means „start at the very bottom level and go up from there” (with public/ being the very bottom level).
That link on cats.html would need to be written as either ./dogs.html (start one directory back and work up) or /animals/dogs.html (explicitly state which directory to start at).
Absolute URLs get longer, naturally, the more complex the directory structure.
With a structure like this, for dogs.html to link to cats.html, it would have to be either…
<!-- Notice the TWO dots, meaning back up another folder level -->
<a href="../c/cats.html">cats</a>
<!-- Or absolute -->
<a href="/animals/pets/c/cats.html">cats</a>
It’s worth noting in this scenario that if animals/ was renamed animal/, then the relative link would still work, but the absolute link would not. That can be a downside to using absolute links. When you’re that specific, making a change to the path will impact your links.
We’ve only looked at HTML linking to HTML, but this idea is universal to the web (and computers, basically). For example, in a CSS file, you might have:
body {
/* Back up one level from /images and follow this path */
background-image: url(./images/pattern.png);
}
…then that becomes wrong because your CSS file is now nested in another directory and is referencing paths from a deeper level. You’d need to back up another folder level again with two dots, like ../images/pattern.png.
One URL format isn’t better than another — it just depends on what you think is more useful and intuitive at the time.
For me, I think about what is the least likely thing to change. For something like an image asset, I find it very unlikely that I’ll ever move it, so linking to it with an absolute URL path (e.g. /images/pattern.png) seems the safest. But for linking documents together that all happen to be in the same directory, it seems safer to link them relatively.
Guides, resources and discussions about Semantic HTML are often focused around specific elements, like a heading, or a sectioning element, or a list. It’s not often that we talk specifically about how we can combine HTML elements to increase their effectiveness.
Normally, when we introduce HTML, we talk about how it is used to apply meaning to content in a document, and we do this by using examples like:
„Is it a paragraph?”
„Is it a heading?”
„Is it a numbered list”
„Is it a button”
We use these examples because they are easy to understand — it’s a single piece or chunk of the same content that needs to be represented in a particular way. This is useful, but it only scratches the surface of how we can use and combine elements to provide more context and meaning.
You remember Voltron, right? Each member of the Voltron force was powerful in their own right, but it was when they combined together to form a towering figure that their mighty powers became unstoppable.
The same is true of HTML elements. I have a few favorite combinations that I’ll take you through. They may seem obscure, but you’d be surprised at how often they come up when you take the time to think outside of divs and spans.
Abbreviations and Definitions
<abbr> and <dfn> are two of my favorite HTML elements. I particularly like them because they work really well together.
You can combine the abbreviation and definition elements to allow browsers, search bots, and other technologies to recognize that something is being defined and that the acronym is associated to that phrase.
<p>
The <dfn><abbr title="International Good
Dog Association">IGDA</abbr></dfn> is an international,
not-for-profit organization responsible for determining
that all dogs are good.
</p>
In the above example, I’m defining that the acronym „IGDA” as „International Good Dogs Association.” I do this by wrapping the acronym in an <abbr> element with a title attribute defining the full name. By adding the <dfn> element around the abbreviation, it indicates that the surrounding paragraph defines the term „International Good Dogs Association.”
The <abbr> element is useful because it can tell screen readers what they should read, while also providing a useful visual representation in the form of a tooltip explaining what the abbreviation is.
Visual representation of <abbr> and <dfn>.
Keyboard, Sample and Variable
If you haven’t heard of these elements, then get ready to have your socks blown off, because they are awesome.
First up, the <kbd> element is used to represent text for a textual user input (e.g. from a keyboard). You can also nest multiple <kbd> elements to represent multiple keystrokes. I love this because, as developers, we find ourselves (hopefully) writing documentation, blog posts, and guides on a regular basis and HTML provides us with native ways to represent this content straight out of the box!
If I wanted to tell someone how you copy and paste with the keyboard, I could mark it up like the code below.
<p>I like to <kbd><kbd>Ctrl</kbd>+<kbd>C</kbd></kbd> and <kbd><kbd>Ctrl</kbd>+<kbd>V</kbd></kbd> a lot.</p>
It looks a bit nuts but the above code, when rendered, looks like the following without any styling applied to it. If you are wondering why kbd is nested inside another kbd element, what this does is specify that the key or input is part of a larger input. Even more combined superpowers!
You can further target the <kbd> elements with CSS to make it look more like little keyboard controls. (Note: Browsers tend to render this by default with a monospace font.)
If you’re wondering what the difference is between using <kbd> versus a span, I believe it comes down to information. I will repeat this sentiment a lot: we do not know how someone is going to consume our HTML, so give your content it’s best chance by representing it in the most meaningful and contextual way possible. If you are still not on board, then please go read my post about HTML as told by TypeScript.
The <samp> element is really cool because you can nest it inside the <kbd> element and vice versa. WHAT? I know, so versatile! Let’s have a look at some examples from MDN.
The following code is an example of nesting a <samp> element inside a <kbd> element. This is used to mark up content that represents input based on the text displayed by the system (e.g. button names).
<p>To save the image file, select <kbd><kbd><samp>File</samp></kbd> - <kbd><samp>Save as...</samp></kbd></kbd>.
</p>
In the above code, we define our keyboard shortcuts the same as our previous example, but we also determine that the menu and menu item names (contained within both <kbd> and <samp>) are an input selected from something displayed by the system, e.g. a dialog or widget.
In other words, this piece of text is a message from the system which has some user inputs that you need to follow (like File and Save as…).
Whereas, when we nest <kbd> inside <samp>, we determine that the input has been echoed back to the user by the system.
<p><samp>yarn start:theproject does not exist, did you mean:</p>
<blockquote><samp><kbd>yarn start:the-project</kbd></samp></blockquote>
Finally, the <var> element! This is used to mark up the name of a variable in math or programming, for example:
<var>E</var> = <var>m</var><var>c</var><sup>2</sup>.
<samp>Error: <var>console<var> is undefined.</samp>
Here you can start to see how combining <var> with other elements like <pre>, <code>, <kbd> or <samp> starts to make your content’s markup more explicit by adding more context. Anything that interprets your HTML markup can start to derive more meaning from the elements you are using rather than just assuming that it’s all standard text.
If you put this content in a paragraph with some spans, there is no way for technology to distinguish this from any other old text on your page. You don’t have to resort to <span> or a <div> to represent this content because HTML already provides us with more semantically accurate elements we can use. HTML is not just about presentation; it’s about meaning. Various technologies outside of visual rendering engines rely on this meaning to make decisions about how to communicate our content in the most meaningful way (e.g. screen readers, text to voice, reading apps, bots, or the next big thing in the future).
Figures
Figures (<figure>) are a great example of a power combination element. Unfortunately, I think it is widely misused and under appreciated (much like <aside>, which I could talk about for hours). The obvious combination you are probably familiar with is using <figure> and <figcaption> together. We often use this duo to represent graphical content like images, illustrations and diagrams — but it can also be used for things like code, poems and quotes!
Because a figure is so flexible in what kind of content it represents, there are a bunch of different HTML elements you can use within a figure to provide more context around the type of content you are putting in your page.
Figcaption
The <figure> element is often seen with an optional <figcaption> which represents a caption or legend for a figure. It should be the first child or last child of the <figure> element. You can also use any flow content (e.g. paragraphs, headings, etc.) in the <figcaption> to provide more context and you can have multiple images inside a <figure> represented by a single <figcaption>.
<figure>
<img src="jello.jpg" alt="Golden Retriever Puppy" />
<figcaption>Jello the Golden Retriever enjoying being carried home.</figcaption>
</figure>
Preformatted Text
The <pre> element is used to display preformatted text or code and is usually rendered using a monospace font. While <pre> can be used on its own, it can also be combined with <figure> and <figcaption>. By doing this, the content inside the <pre> element becomes more accessible to assistive technologies, like screen readers, since it allows us to use <figcaption> as a label for the text or code.
We can go even further by combining the <pre> and <code> elements to identify the pre- formatted content as code.
Because <code> is considered flow content, it can also be used inside a <figcaption> to mark up inline references to code (i.e. a single phrase or line).
Cite and Blockquote
Quotations are something I use a lot and it honestly didn’t occur to me it could be used as part of a figure. But it still makes sense to use cite or quotation if the content they contain is relevant to the overall content, even if it is not part of the main document flow.
I like the example of using a figure for poems, mainly so I could share the amazing poem I wrote about my dog. Here I use the <cite> element inside the <figcaption>:
<figure>
<p>
Jello - A little fluffy dog.
Hello!
Squishy Jello - you little fluffy fellow.
BOUNCY yellow Jello.
A very mellow fellow.
EATING MY MARSHMALLOW.
</p>
<figcaption><cite>Eating my Marshmallow by Mandy Michael</figcaption>
</figure>
It’s easy to fall into the trap of thinking that the figure element is just for images or image-like content, but you can use it for content like audio, video, charts, poems, quotations or even tables of statistics. Because the element is so versatile you can combine it with so many other elements to provide more and more context about the figure for assistive technology, browsers, and other technologies consuming your website.
Wrapping Up
These are just some of the ways HTML elements can be combined for better context. HTML elements are indeed useful and valuable on their own when used as isolated pieces — and using them this way is a good first step. But, like Voltron, when you combine HTML elements together, the individual pieces form a greater whole and gain much more meaning and power. It’s important that we don’t think of HTML as individual pieces of code, but parts of a whole because HTML is really a mass of totally amazing combinations.
You can combine and use HTML elements in any number of ways to best represent your content. Don’t simply stick to the same old things you know; take the time to explore HTML and learn about all it has to offer. Like any language, we should make the most of it and use it to its full potential.
The more accurate you are in marking up your content, the better that content will be represented across technologies, and the more prepared it will be for any anything else that is used to interpret your HTML in the future.
So, go forth and HTML elements, unite!
Resources
In my opinion, the best resource for learning and understanding how to use HTML (outside of the spec itself) is the MDN Web Docs. It lists all the HTML elements you could ever need.
The following are among the elements covered in this post:
We’ve all seen articles like „The Top 5 Ways To Fix Your Sign Up Flow and Get On With Your Life.” Articles like this aren’t wrong or bad, they are just shallow and a bit junk food-y and BuzzFeed-y. Of course, a designer’s actual job is complicated, nuanced, and difficult. But deep dives into all that are far less common.
It’s clear that the currency of design discourse is really concerned with the “how” of design, not the “why” of it. As Teixeira and Braga write:
While designers tend to be skeptical of magic formulas—we’re decidedly suspicious of self-help gurus, magic diets, or miraculous career advice—we have a surprisingly high tolerance for formulaic solutions when it comes to design.
That’s a pointed criticism but, from my perspective, it’s also quite accurate.
There’s not much talk about frameworks here. There’s no shaming about old techniques, or jokes about JavaScript. There’s just a couple hundred people all around me laughing and smiling and watching talks about making things on the web and it all feels so fresh and new to me. Unlike many other conferences I’ve visited, these talks are somehow inclusive and rather feel, well, there’s no other word for it: inspiring.
I’m sitting in a little room buried underneath the Veterans Memorial Coliseum in Portland and I’m here for my third XOXO. And I can’t stop smiling.
Although the festival is not entirely focused on coding and front-end development, there are a lot of developers here that make art on the web for fun. From Jenn Schiffer’s pixel art to Monica Dinculescu’s emoji projects and Nicole He’s buck-wild enhance.computer, there’s a lot of interesting discussions about coding — but! — it’s from a very different perspective than the one I’m familiar with.
Most conferences tend to focus on being practical. Here’s the newest technique! Here’s how to improve your career! Here’s the coolest new folks that you should be following! But it’s important to remember that the web isn’t only a serious place for serious work. It can be this entirely other thing, too.
The web can be for fun. It can be utterly weird and unexpected. And that’s what we’re all seeing in this little room right now at XOXO; websites that can’t be monetized, websites that can’t be controlled by corporate interests or giant ad networks.
I got a great email from a fellow named Josh Long the other day. He is, in his words, „relatively new to web design” and was a bit stuck on the concept of getting a site live. I should say that I’m happy to get emails like this an I always read them, but I typically can’t offer tech support over email. If I can respond at all, I normally point people to other community resources.
In this case, it struck me what a perfect moment this is for Josh. He’s a little confused, but he knows enough to be asking a lot of questions and sorting through all this stuff. I figured this was a wonderful opportunity to dig into his questions, hopefully helping him and just maybe helping others in a similar situation.
Here’s one of the original paragraphs Josh sent me, completely unedited:
I’m relatively new to web design, but I’ve taken a few courses on HTML and CSS and I’ve done a Codecademy course on JavaScript. But, (jumping forward probably quite a while here!) after having fully designed and coded a HTML/CSS/JS website or webpage, I don’t fully understand the full process of going from a local site hosted with mamp/wamp to publishing a public site using wordpress(?) or some other host (is WordPress even a host?!) and also finding a server that’s suitable and some way of hosting images/videos or other content. (If that sounded like I didn’t know what half of those meant, it’s because unfortunately I don’t!.. but I’d really like to!)
Can you sense that enthusiasm? I love it.
We worked together a bit to refine many of his questions for this post, but they are still Josh’s words. Here we go!
What is a Domain Registrar? I get they are for registering domain names, but what’s the difference between them? How do you know which one is right for you? A quick search for „best domain hosts” on Google gave me 5 ads for companies who are domain registrars/hosts and 9 „Top 10” style pages that look as though they have some sort of affiliation with at least one company they’re suggesting. Am I just looking for the cheapest one?
You’re exactly right, domain registrants are for registering domain names. If you want joshlongisverycool.com, you’re going to have to buy it, and domain registrants are companies that help you do that.
Searching for a domain name
It’s a bummer that web search results are so saturated by ads and affiliate-link-saturated SEO-jacked pages. You’ll never get total honesty from that because those pages are full of links promoting whoever will pay them the most to send new customers. Heck, even Google themselves will sell you a domain name.
The truth is that you can’t go too wrong here. Domain names are a bit of a commodity and the hundreds of companies that will sell you one largely compete on marketing.
Some things to watch out for:
Some companies treat the domain as a loss leader. Like a grocery store selling cheap milk to hope you buy some more stuff while you are there. The check out process at any domain registrant will almost certainly try to sell you a bunch of extra stuff. For example, they might try to sell you additional domain names or email hosting you probably don’t need. Just be careful.
Web hosts (which we’re getting to next) will often sell them to you along with hosting. That’s fine I suppose, but I consider it a bit of a conflict of interest. Say you choose to move hosts one day. That hosting company is incentivized in the wrong direction to make that easy for you. If the domain is handled elsewhere, that domain registrant is incentivized the right direction to help you make changes.
I hate to add to the noise for you, but here are some domain registrants that I’ve used personally and aren’t paying for sponsorship here nor are affiliate links:
Our own Sarah Drasner recommends looking at ZEIT domains, which are super interesting in that you buy and manage them entirely over the command line.
I might suggest, if you can see yourself owning several domain names in your life, keeping them consolidated to a single registrant. Managing domains isn’t something you’ll do very often, so it’s easy to lose track of what domains you registered on what registrant, not to mention how/where to change all the settings across different registrants.
I’d also suggest it’s OK to experiment here. That’s how all of us have learned. Pick one with a UI that you don’t hate and a trustable vibe. Maybe your friend also uses that one. Hopefully, it works out fine. If you hate it, it’ll be a little work, but you can always switch.
What is a web host and why do I need one? A Google search throws you a mountain of „best web host” articles and ads. These websites all seem to be full of jargon like „shared hosting” and „managed hosting.” I see things like „suggested hosts” on some sites. How do you find the right web host? I’m not even sure what my needs are. Should I just find the cheapest one?
Just because you own a domain doesn’t mean it will do anything. In fact, right after you buy it, it’s likely that the domain registrant slaps up a „coming soon” page for you:
A „coming soon” page you might see immediately after buying a domain name.
To host a website at your new domain, you’ll need to configure the DNS of your new domain to point at a server connected to the internet. Here’s an interesting tidbit… you could do this right from your house if you wanted to. Your Internet Service Provider (ISP) at home probably gives you an IP address. It’s all a bit nerdy, but you could point your domain name at that IP and set up your computer to be a web server that responds to incoming requests and serves up your website. Almost nobody does that though. You don’t want your web server to stop working because you closed your laptop or your ISP changed your IP.
Web hosting services give you that server. Like domain registrants, web hosts are almost a commodity. There are lots of them that provide a similar service, so the price is driven fairly low and they find other things to compete on.
Buying web hosting is a little trickier than buying a domain though. You’ll need to know a little bit about the website you intend to host when making that choice. Will it be a WordPress site? Or another PHP/MySQL-based CMS site (we’ll get to those later)? That means your host will need to support those technologies. Most do, some don’t. You might wanna look in their documentation or literally ask them before pulling the trigger if you are unsure. There are lots of technologies for running websites. Say the site will use Ruby on Rails — that’s a different set of requirements that not all hosts offer. Or Node… Or Python… same story.
If a web host says they specialize in a particular set of technologies, and that’s what you need, that’s probably a decent way to go, particularly in your early days. Let’s take a very limited gander. Again, these are not affiliate or paid-for links and they are somewhat randomly selected web hosts that come to mind for me:
WP Engine is a web host that focuses specifically on WordPress.
BlueHost is has very inexpensive hosting choices, but has essentially the same capabilities as those others.
Now here some other web hosts that are a little less traditional. Forgive the techy terms here — if they don’t mean anything to you, just ignore them.
Netlify does static site hosting, which is great for things like static site generators and JAMstack sites.
Zeit is a host where you interact with it only through the command line.
Digital Ocean has their own way of talking about hosting. They call their servers Droplets, which are kind of like virtual machines with extra features.
Heroku is great for hosting apps with a ready-to-use backend for things like Node, Ruby, Java, and Python.
Again, I’d say it’s OK to, in a sense, make some mistakes here. If you aren’t hosting something particularly mission-critical, like your personal website, pick a host that seems to fit the bill and give it a go. Hopefully, it works out great — if not, you can move. Moving isn’t always super fun, but everybody ends up doing it, and you’ll learn as you go.
When you buy web hosting, that host is going to tell you how to use it. One common way is that the host will give you SFTP credentials. Then you’ll use software that is built for connecting to servers over SFTP, and that will allow you to upload files to the web server.
This is a magic moment!
Say you’ve been working on a file that is a single index.html file, that loads a style.css file. Upload those files over SFTP into the folder your host tells you is correct public root directory for that site.
That is the process for taking a site from local to live! There is nothing wrong with it, either. This is referred to as deployment, and this is about as basic and simple as it gets. Even fancier ways of doing deployment sometimes ultimately do just this behind the scenes. We’ll get more into deployment later.
Should you bundle your domain registrar and web host into one if a company offers both?
I mentioned this above a little: I’m a fan of not doing that in general. On one hand, it’s mighty handy. Things like shared billing and a single checkout flow. The host will also do things like configuring the DNS for you to be all set up for their hosting and you probably don’t even have to think about it.
But say the day comes where you just don’t like that host anymore. You’ve found a better deal, outgrown them, were turned off by their support or service, or any other reason. You want to move hosts. The problem is that they aren’t just your host, but your domain registrant, too. Do you leave the domain with them and just move hosts? Probably not, you’re trying to leave them. Now you need to move two things. That makes that move all the more perilous, but worse, this company isn’t exactly incentivized to respond quickly and helpfully to your requests since they know they’re losing you as a customer.
What really is a „CMS”? What’s its purpose? WordPress, Joomla, and Drupal are the most popular names I’ve found for content management systems, and from their descriptions, all sound very similar. What are the kinds of features that set one apart from another? Is a CMS all you need to get your website from your local computer to the public internet?
CMS (Content Management System) is a pretty generic term. It means literally anything that helps you manage content. There are, as you’ve seen, some big players like WordPress and CraftCMS. They don’t really have anything directly to do with that connection between working locally on a site and getting that site live. But they do rather complicate it.
The reason that you’d use a CMS at all is to make working on your site easier. Consider this site you’re looking at right now. There are tens of thousands of pages on this site. It would be untenable for each of them to be a hand-authored file.html file.
Instead, a CMS allows us to craft all those pages by combining data and templates.
Let’s consider the technology behind WordPress, a CMS that works pretty good for CSS-Tricks. In order for WordPress to run, it needs:
PHP (the back-end language)
MySQL (the database)
Apache (the web server)
You can do all that locally!
I use Local by Flywheel for that (Mac and Windows), but there are a number of ways to get those technologies running: MAMP, Docker, Vagrant, etc.
That’ll get you up and running locally with your CMS. That’s a beautiful thing. Having a good local development environment for your site is crucial. But it doesn’t help you get that site live.
We’ll go over getting all this to a live site a bit later, but you should know this: the web server that you run will need to run these same technologies. Looking at the WordPress example from above, your web server will also need to run PHP, MySQL, and Apache. Then you’ll need to set up a system for getting all your files from your local computer to your web server like you would for any other site, but also probably have a system for dealing with the database. The database is a bit tricky as it’s not a „flat file” like most of the rest of your site.
A CMS could be built from any set of technologies, not just the ones listed above. For example, see KeystoneJS. Instead of PHP, Keystone is Node.js. Instead of MySQL for the database, it uses MongoDB. Instead of Apache, it uses Express. Just a different set of technologies. Both of which you can get running locally and on a live web server.
A CMS could even have no database at all! Static site generators are like this. You get the site running locally, and they produce a set of flat files which you move to your live server. A different way to do things, but absolutely still a CMS. What I always say is that the best CMS is one that is customized to your needs.
What is „Asset Hosting”? Are assets not content? What is the difference between a CMS and an asset hosting service? What does an asset host do?
Let’s define an asset: any „flat” file. As in, not dynamically generated like how a CMS might generate an HTML file. Images are a prime example of an „asset.” But it’s also things like CSS and JavaScript files, as well as videos and PDFs.
And before we get any further: you probably don’t need to worry about this right away. Your web host can host assets and that’s plenty fine for your early days on small sites.
One major reason people go with an asset host (probably more commonly referred to as a CDN or Content Delivery Network) is for a speed boost. Asset hosts are also servers, just like your web host’s web server, but they are designed for hosting those flat file assets super fast. So, not only do those assets get delivered to people looking at your site super fast, but your web server is relieved of that burden.
You could even think of something like YouTube as an asset host. That 100 MB video of a butterfly in your garden is a heavy load for your little web server, and potentially a problem if your outgoing bandwidth is capped like it often is. Uploading that video to YouTube puts your video into that whole social universe, but a big reason to do it other than the social stuff is that it’s hosting that video asset for you.
I’ve heard of „repositories”, but don’t really get what they are. I hear stuff like „just upload it to my Git Repository.” What the heck does that mean? I feel like a moron for asking this. What is Git? What is it for? Do I need to use it? Is it involved in the „local to live” process at all?
Sorry you got hit with a „just” there. There is an epidemic in technology conversations where people slip that in to make it seem like what they are about to say is easy and obvious when it could be neither depending on who is reading.
But let’s get into talking about Git repositories.
Git is a specific form of version control. There are others, but Git is so dominant in the web industry that it’s hardly worth mentioning any others.
Let’s say you and me are working on a website together. We’ve purchased a domain and hosting and gotten the site live. We’ve shared the SFTP credentials so we both have access to change the files on the live site. This could be quite dangerous!
Coda is a code editor that let’s you edit files directly on a sever you’ve connected to over SFTP. Cool, but dangerous!
Say we both edit a file and upload it over SFTP… which change wins? It’s whoever uploaded their file last. We have no idea who has done what. We’re over-writing each other and have no way of staying in sync with each other’s changes, changes immediately affect the live site which could break things and we have no way of undoing changes in case we break things. That’s a situation so unacceptable that it’s really never done.
Instead, we work with version control, like Git. With Git, when we make changes, we commit them to a repository. A repository can be hosted anywhere, even on your local machine. But to make them really useful, they are hosted on the internet somewhere everyone has access to. You’ve surely seen GitHub, which hosts these repositories and adds a bunch of other features like issue tracking. Similar are GitLab and Bitbucket.
Now let’s say you and me are working on that same site, but we’ve set up a Git repository for it. When I make a change, I commit it to the repository. If you want to make a change as well, you have to pull my changes down which merges them into your own copy of the code. Then you can push your changes up to the repository. Like anything, it gets more complicated, but that’s the gist of it.
But a Git repository isn’t the live website. You’re on your own for getting the files from a Git repository to a live site. Fortunately, that’s a situation that everyone faces, so there are lots of options. Good thing your last question is about this!
OK. So now with all that straight… where do you start with going from local to live? Where do you „upload” your HTML, CSS and JavaScript files? How do you link your shiny new domain name to those files and see it out in the wild? Which service is in charge of adding new content to your site, or updating it? Does it get really confusing if you have different companies for each service?
Let’s start with a very simple website on what I’d consider typical web hosting. Say you have just index.html, style.css, and script.js files on your local computer which is your entire website. You’ve purchased a domain name and pointed the DNS settings at a web host. That host has given you SFTP credentials. You’ll use those credentials in an app that allows SFTP connections to log in:
Your host will also tell you which folder is the „public root” of your website. The files there will be out on the public internet for the world to see!
You might hear people refer to the „live” website as a „production” site. When someone asks something like, Did that bug make it to production?” they mean whether the bug is on the live website. „Development” is your local computer. You might also have a „Staging” site, which is a clone of the live website on the same hardware/software of the live site for testing.
Remember earlier when we talked about Git repositories? While the repositories themselves don’t directly help you get the files in them to your web server, most systems that help you with the local-to-live process work with your repositories.
The phrase „local-to-live” refers to deployment. When you have changes that you want to go out to your production website, you deploy them. That’s the process of moving your work from „development” to „production.”
One service that helps with this idea of deployment is Beanstalk. Beanstalk hosts your Git repository, plus you give it the SFTP credentials for your server — that way it can move the files to your web server when you make commits. Cool right? Say you wanted to host that Git repo elsewhere though, like GitHub, Bitbucket, or Gitlab. Check out DeployBoy, which does the same thing, only it can connect to those sites as well. It probably comes as no surprise by now that there are lots of options here, ranging in price and complexity.
Let’s go back to our WordPress example.
You’ve got it running locally (on your computer) just perfectly and now want to go live with it.
You’ve bought a domain name from a registrar.
You’ve purchased hosting that meets WordPress requirements.
You’ve pointed the DNS of the domain name at the web host.
You’ve verified it’s all working (easy way: upload an index.html file in the public root via SFTP and verify that it loads when you type you type the domain name into a browser.)
Now, you’ve still got some work to do:
Set up a Git repository for the site.
Set up a deployment service to move the files from the repository to the live site.
Configure/Set up the live site as needed. For example, you’ll need a database on the live site. You’ll have to create that (your host will have instructions) and do things like run the WordPress installer and update configuration files.
If you have things in your local database that you want to move to the live site, you might be exporting/importing things. That can happen at the raw MySQL level using WordPress’ native import/export features, or a fancy plugin like WP DB Migrate Pro.
It’s a non-trivial amount of work! Sorry. This process is pretty similar for any site though. It’s a matter of configuring and setting up your production web server exactly how it should be, then deploying files to it. Every site is a bit different, but you’ll get the hang of this whole dance.
It really is a big dance. I’ve only painted one picture for you here. I tried to pick one that was generic and broad enough that it shows the landscape of what needs to be done. But, at every step in this dance, there are different ways you can do things, different services you can pick, companies trying to help you at different pain points… it’s an ever-changing world.
Right now, Netlify is enjoying a lot of popularity because they are one of the only web hosts that actually helps you with deployment. They’ll watch your Git repositories and do deployment for you! Netlify is for static sites only, but that can be a whole world onto itself. ZEIT also is massively innovative in how it helps with deploying and hosting web projects, including directly connecting with GitHub.
Good luck!
I hope this was helpful. Remember, you aren’t alone in all this. Zillions of other developers have done this before you and there is help to be found on the internet.
Oh, and remember: the best way to learn anything at all is to…
I got a great email from a fellow named Josh Long the other day. He is, in his words, „relatively new to web design” and was a bit stuck on the concept of getting a site live. I should say that I’m happy to get emails like this an I always read them, but I typically can’t offer tech support over email. If I can respond at all, I normally point people to other community resources.
In this case, it struck me what a perfect moment this is for Josh. He’s a little confused, but he knows enough to be asking a lot of questions and sorting through all this stuff. I figured this was a wonderful opportunity to dig into his questions, hopefully helping him and just maybe helping others in a similar situation.
Here’s one of the original paragraphs Josh sent me, completely unedited:
I’m relatively new to web design, but I’ve taken a few courses on HTML and CSS and I’ve done a Codecademy course on JavaScript. But, (jumping forward probably quite a while here!) after having fully designed and coded a HTML/CSS/JS website or webpage, I don’t fully understand the full process of going from a local site hosted with mamp/wamp to publishing a public site using wordpress(?) or some other host (is WordPress even a host?!) and also finding a server that’s suitable and some way of hosting images/videos or other content. (If that sounded like I didn’t know what half of those meant, it’s because unfortunately I don’t!.. but I’d really like to!)
Can you sense that enthusiasm? I love it.
We worked together a bit to refine many of his questions for this post, but they are still Josh’s words. Here we go!
What is a Domain Registrar? I get they are for registering domain names, but what’s the difference between them? How do you know which one is right for you? A quick search for „best domain hosts” on Google gave me 5 ads for companies who are domain registrars/hosts and 9 „Top 10” style pages that look as though they have some sort of affiliation with at least one company they’re suggesting. Am I just looking for the cheapest one?
You’re exactly right, domain registrants are for registering domain names. If you want joshlongisverycool.com, you’re going to have to buy it, and domain registrants are companies that help you do that.
Searching for a domain name
It’s a bummer that web search results are so saturated by ads and affiliate-link-saturated SEO-jacked pages. You’ll never get total honesty from that because those pages are full of links promoting whoever will pay them the most to send new customers. Heck, even Google themselves will sell you a domain name.
The truth is that you can’t go too wrong here. Domain names are a bit of a commodity and the hundreds of companies that will sell you one largely compete on marketing.
Some things to watch out for:
Some companies treat the domain as a loss leader. Like a grocery store selling cheap milk to hope you buy some more stuff while you are there. The check out process at any domain registrant will almost certainly try to sell you a bunch of extra stuff. For example, they might try to sell you additional domain names or email hosting you probably don’t need. Just be careful.
Web hosts (which we’re getting to next) will often sell them to you along with hosting. That’s fine I suppose, but I consider it a bit of a conflict of interest. Say you choose to move hosts one day. That hosting company is incentivized in the wrong direction to make that easy for you. If the domain is handled elsewhere, that domain registrant is incentivized the right direction to help you make changes.
I hate to add to the noise for you, but here are some domain registrants that I’ve used personally and aren’t paying for sponsorship here nor are affiliate links:
Our own Sarah Drasner recommends looking at ZEIT domains, which are super interesting in that you buy and manage them entirely over the command line.
I might suggest, if you can see yourself owning several domain names in your life, keeping them consolidated to a single registrant. Managing domains isn’t something you’ll do very often, so it’s easy to lose track of what domains you registered on what registrant, not to mention how/where to change all the settings across different registrants.
I’d also suggest it’s OK to experiment here. That’s how all of us have learned. Pick one with a UI that you don’t hate and a trustable vibe. Maybe your friend also uses that one. Hopefully, it works out fine. If you hate it, it’ll be a little work, but you can always switch.
What is a web host and why do I need one? A Google search throws you a mountain of „best web host” articles and ads. These websites all seem to be full of jargon like „shared hosting” and „managed hosting.” I see things like „suggested hosts” on some sites. How do you find the right web host? I’m not even sure what my needs are. Should I just find the cheapest one?
Just because you own a domain doesn’t mean it will do anything. In fact, right after you buy it, it’s likely that the domain registrant slaps up a „coming soon” page for you:
A „coming soon” page you might see immediately after buying a domain name.
To host a website at your new domain, you’ll need to configure the DNS of your new domain to point at a server connected to the internet. Here’s an interesting tidbit… you could do this right from your house if you wanted to. Your Internet Service Provider (ISP) at home probably gives you an IP address. It’s all a bit nerdy, but you could point your domain name at that IP and set up your computer to be a web server that responds to incoming requests and serves up your website. Almost nobody does that though. You don’t want your web server to stop working because you closed your laptop or your ISP changed your IP.
Web hosting services give you that server. Like domain registrants, web hosts are almost a commodity. There are lots of them that provide a similar service, so the price is driven fairly low and they find other things to compete on.
Buying web hosting is a little trickier than buying a domain though. You’ll need to know a little bit about the website you intend to host when making that choice. Will it be a WordPress site? Or another PHP/MySQL-based CMS site (we’ll get to those later)? That means your host will need to support those technologies. Most do, some don’t. You might wanna look in their documentation or literally ask them before pulling the trigger if you are unsure. There are lots of technologies for running websites. Say the site will use Ruby on Rails — that’s a different set of requirements that not all hosts offer. Or Node… Or Python… same story.
If a web host says they specialize in a particular set of technologies, and that’s what you need, that’s probably a decent way to go, particularly in your early days. Let’s take a very limited gander. Again, these are not affiliate or paid-for links and they are somewhat randomly selected web hosts that come to mind for me:
WP Engine is a web host that focuses specifically on WordPress.
BlueHost is has very inexpensive hosting choices, but has essentially the same capabilities as those others.
Now here some other web hosts that are a little less traditional. Forgive the techy terms here — if they don’t mean anything to you, just ignore them.
Netlify does static site hosting, which is great for things like static site generators and JAMstack sites.
Zeit is a host where you interact with it only through the command line.
Digital Ocean has their own way of talking about hosting. They call their servers Droplets, which are kind of like virtual machines with extra features.
Heroku is great for hosting apps with a ready-to-use backend for things like Node, Ruby, Java, and Python.
Again, I’d say it’s OK to, in a sense, make some mistakes here. If you aren’t hosting something particularly mission-critical, like your personal website, pick a host that seems to fit the bill and give it a go. Hopefully, it works out great — if not, you can move. Moving isn’t always super fun, but everybody ends up doing it, and you’ll learn as you go.
When you buy web hosting, that host is going to tell you how to use it. One common way is that the host will give you SFTP credentials. Then you’ll use software that is built for connecting to servers over SFTP, and that will allow you to upload files to the web server.
This is a magic moment!
Say you’ve been working on a file that is a single index.html file, that loads a style.css file. Upload those files over SFTP into the folder your host tells you is correct public root directory for that site.
That is the process for taking a site from local to live! There is nothing wrong with it, either. This is referred to as deployment, and this is about as basic and simple as it gets. Even fancier ways of doing deployment sometimes ultimately do just this behind the scenes. We’ll get more into deployment later.
Should you bundle your domain registrar and web host into one if a company offers both?
I mentioned this above a little: I’m a fan of not doing that in general. On one hand, it’s mighty handy. Things like shared billing and a single checkout flow. The host will also do things like configuring the DNS for you to be all set up for their hosting and you probably don’t even have to think about it.
But say the day comes where you just don’t like that host anymore. You’ve found a better deal, outgrown them, were turned off by their support or service, or any other reason. You want to move hosts. The problem is that they aren’t just your host, but your domain registrant, too. Do you leave the domain with them and just move hosts? Probably not, you’re trying to leave them. Now you need to move two things. That makes that move all the more perilous, but worse, this company isn’t exactly incentivized to respond quickly and helpfully to your requests since they know they’re losing you as a customer.
What really is a „CMS”? What’s its purpose? WordPress, Joomla, and Drupal are the most popular names I’ve found for content management systems, and from their descriptions, all sound very similar. What are the kinds of features that set one apart from another? Is a CMS all you need to get your website from your local computer to the public internet?
CMS (Content Management System) is a pretty generic term. It means literally anything that helps you manage content. There are, as you’ve seen, some big players like WordPress and CraftCMS. They don’t really have anything directly to do with that connection between working locally on a site and getting that site live. But they do rather complicate it.
The reason that you’d use a CMS at all is to make working on your site easier. Consider this site you’re looking at right now. There are tens of thousands of pages on this site. It would be untenable for each of them to be a hand-authored file.html file.
Instead, a CMS allows us to craft all those pages by combining data and templates.
Let’s consider the technology behind WordPress, a CMS that works pretty good for CSS-Tricks. In order for WordPress to run, it needs:
PHP (the back-end language)
MySQL (the database)
Apache (the web server)
You can do all that locally!
I use Local by Flywheel for that (Mac and Windows), but there are a number of ways to get those technologies running: MAMP, Docker, Vagrant, etc.
That’ll get you up and running locally with your CMS. That’s a beautiful thing. Having a good local development environment for your site is crucial. But it doesn’t help you get that site live.
We’ll go over getting all this to a live site a bit later, but you should know this: the web server that you run will need to run these same technologies. Looking at the WordPress example from above, your web server will also need to run PHP, MySQL, and Apache. Then you’ll need to set up a system for getting all your files from your local computer to your web server like you would for any other site, but also probably have a system for dealing with the database. The database is a bit tricky as it’s not a „flat file” like most of the rest of your site.
A CMS could be built from any set of technologies, not just the ones listed above. For example, see KeystoneJS. Instead of PHP, Keystone is Node.js. Instead of MySQL for the database, it uses MongoDB. Instead of Apache, it uses Express. Just a different set of technologies. Both of which you can get running locally and on a live web server.
A CMS could even have no database at all! Static site generators are like this. You get the site running locally, and they produce a set of flat files which you move to your live server. A different way to do things, but absolutely still a CMS. What I always say is that the best CMS is one that is customized to your needs.
What is „Asset Hosting”? Are assets not content? What is the difference between a CMS and an asset hosting service? What does an asset host do?
Let’s define an asset: any „flat” file. As in, not dynamically generated like how a CMS might generate an HTML file. Images are a prime example of an „asset.” But it’s also things like CSS and JavaScript files, as well as videos and PDFs.
And before we get any further: you probably don’t need to worry about this right away. Your web host can host assets and that’s plenty fine for your early days on small sites.
One major reason people go with an asset host (probably more commonly referred to as a CDN or Content Delivery Network) is for a speed boost. Asset hosts are also servers, just like your web host’s web server, but they are designed for hosting those flat file assets super fast. So, not only do those assets get delivered to people looking at your site super fast, but your web server is relieved of that burden.
You could even think of something like YouTube as an asset host. That 100 MB video of a butterfly in your garden is a heavy load for your little web server, and potentially a problem if your outgoing bandwidth is capped like it often is. Uploading that video to YouTube puts your video into that whole social universe, but a big reason to do it other than the social stuff is that it’s hosting that video asset for you.
I’ve heard of „repositories”, but don’t really get what they are. I hear stuff like „just upload it to my Git Repository.” What the heck does that mean? I feel like a moron for asking this. What is Git? What is it for? Do I need to use it? Is it involved in the „local to live” process at all?
Sorry you got hit with a „just” there. There is an epidemic in technology conversations where people slip that in to make it seem like what they are about to say is easy and obvious when it could be neither depending on who is reading.
But let’s get into talking about Git repositories.
Git is a specific form of version control. There are others, but Git is so dominant in the web industry that it’s hardly worth mentioning any others.
Let’s say you and me are working on a website together. We’ve purchased a domain and hosting and gotten the site live. We’ve shared the SFTP credentials so we both have access to change the files on the live site. This could be quite dangerous!
Coda is a code editor that let’s you edit files directly on a sever you’ve connected to over SFTP. Cool, but dangerous!
Say we both edit a file and upload it over SFTP… which change wins? It’s whoever uploaded their file last. We have no idea who has done what. We’re over-writing each other and have no way of staying in sync with each other’s changes, changes immediately affect the live site which could break things and we have no way of undoing changes in case we break things. That’s a situation so unacceptable that it’s really never done.
Instead, we work with version control, like Git. With Git, when we make changes, we commit them to a repository. A repository can be hosted anywhere, even on your local machine. But to make them really useful, they are hosted on the internet somewhere everyone has access to. You’ve surely seen GitHub, which hosts these repositories and adds a bunch of other features like issue tracking. Similar are GitLab and Bitbucket.
Now let’s say you and me are working on that same site, but we’ve set up a Git repository for it. When I make a change, I commit it to the repository. If you want to make a change as well, you have to pull my changes down which merges them into your own copy of the code. Then you can push your changes up to the repository. Like anything, it gets more complicated, but that’s the gist of it.
But a Git repository isn’t the live website. You’re on your own for getting the files from a Git repository to a live site. Fortunately, that’s a situation that everyone faces, so there are lots of options. Good thing your last question is about this!
OK. So now with all that straight… where do you start with going from local to live? Where do you „upload” your HTML, CSS and JavaScript files? How do you link your shiny new domain name to those files and see it out in the wild? Which service is in charge of adding new content to your site, or updating it? Does it get really confusing if you have different companies for each service?
Let’s start with a very simple website on what I’d consider typical web hosting. Say you have just index.html, style.css, and script.js files on your local computer which is your entire website. You’ve purchased a domain name and pointed the DNS settings at a web host. That host has given you SFTP credentials. You’ll use those credentials in an app that allows SFTP connections to log in:
Your host will also tell you which folder is the „public root” of your website. The files there will be out on the public internet for the world to see!
You might hear people refer to the „live” website as a „production” site. When someone asks something like, Did that bug make it to production?” they mean whether the bug is on the live website. „Development” is your local computer. You might also have a „Staging” site, which is a clone of the live website on the same hardware/software of the live site for testing.
Remember earlier when we talked about Git repositories? While the repositories themselves don’t directly help you get the files in them to your web server, most systems that help you with the local-to-live process work with your repositories.
The phrase „local-to-live” refers to deployment. When you have changes that you want to go out to your production website, you deploy them. That’s the process of moving your work from „development” to „production.”
One service that helps with this idea of deployment is Beanstalk. Beanstalk hosts your Git repository, plus you give it the SFTP credentials for your server — that way it can move the files to your web server when you make commits. Cool right? Say you wanted to host that Git repo elsewhere though, like GitHub, Bitbucket, or Gitlab. Check out DeployBoy, which does the same thing, only it can connect to those sites as well. It probably comes as no surprise by now that there are lots of options here, ranging in price and complexity.
Let’s go back to our WordPress example.
You’ve got it running locally (on your computer) just perfectly and now want to go live with it.
You’ve bought a domain name from a registrar.
You’ve purchased hosting that meets WordPress requirements.
You’ve pointed the DNS of the domain name at the web host.
You’ve verified it’s all working (easy way: upload an index.html file in the public root via SFTP and verify that it loads when you type you type the domain name into a browser.)
Now, you’ve still got some work to do:
Set up a Git repository for the site.
Set up a deployment service to move the files from the repository to the live site.
Configure/Set up the live site as needed. For example, you’ll need a database on the live site. You’ll have to create that (your host will have instructions) and do things like run the WordPress installer and update configuration files.
If you have things in your local database that you want to move to the live site, you might be exporting/importing things. That can happen at the raw MySQL level using WordPress’ native import/export features, or a fancy plugin like WP DB Migrate Pro.
It’s a non-trivial amount of work! Sorry. This process is pretty similar for any site though. It’s a matter of configuring and setting up your production web server exactly how it should be, then deploying files to it. Every site is a bit different, but you’ll get the hang of this whole dance.
It really is a big dance. I’ve only painted one picture for you here. I tried to pick one that was generic and broad enough that it shows the landscape of what needs to be done. But, at every step in this dance, there are different ways you can do things, different services you can pick, companies trying to help you at different pain points… it’s an ever-changing world.
Right now, Netlify is enjoying a lot of popularity because they are one of the only web hosts that actually helps you with deployment. They’ll watch your Git repositories and do deployment for you! Netlify is for static sites only, but that can be a whole world onto itself. ZEIT also is massively innovative in how it helps with deploying and hosting web projects, including directly connecting with GitHub.
Good luck!
I hope this was helpful. Remember, you aren’t alone in all this. Zillions of other developers have done this before you and there is help to be found on the internet.
Oh, and remember: the best way to learn anything at all is to…
Tired of slow, unreliable web hosting? See how easy it is to self-host your next project on DigitalOcean’s cloud platform. Build and manage ultra-fast websites, blogs, and other static web pages using our user-friendly control panel or simple API, all with a 99.99% uptime SLA. Save time using our One-Click install apps for WordPress, Ghost and Discourse. Never worry about running out of storage space again with Spaces — highly scalable, affordable object storage.
Sign up today with a free $100 credit for CSS-Tricks readers.
Most of the behavior in an application revolves around events. User enters a value in the registration form? Event. User hits the submit button? Another event. Events are triggered a number of ways and we build applications to listen for them in order to do something else in response.
You may already be super comfortable working with events based on your existing JavaScript experience. However, React has a distinct way of handling them. Rather than directly targeting DOM events, React wraps them in their own event wrapper. But we’ll get into that.
Let’s go over how to create, add and listen for events in React.
Creating Events
We’ll start by creating a form that has an input and a button. An event will be triggered when a value is entered. The button is used to call a function which will reverse that value.
Here’s how it’ll work:
An empty input field allows the user to enter text.
An onChange event is triggered when values are entered in the input. This calls a function — handleChange() — that is used to set a new state for the input.
When the „Reverse Text” button is clicked, another event is triggered. This calls a function — handleReverse() — to set a new state for reversedText.
Here’s that translated into code:
class App extends React.Component {
state = {
/* Initial State */
input: "",
reversedText: ""
};
/* handleChange() function to set a new state for input */
handleChange = event => {
const value = event.target.value;
this.setState({
input: value
});
};
/* handleReverse() function to reverse the input and set that as new state for reversedText */
handleReverse = event => {
event.preventDefault();
const text = this.state.input;
this.setState({
reversedText: text
.split("")
.reverse()
.join("")
});
};
render() {
return (
<React.Fragment>
{ /* handleReverse() is called when the form is submitted */ }
<form onSubmit={this.handleReverse}>
<div>
{ /* Render input entered */}
<label>Text: {this.state.input}</label>
</div>
<div>
{ /* handleChange() is triggered when text is entered */ }
<input
type="text"
value={this.state.input}
onChange={this.handleChange}
placeholder="Enter a text"
/>
</div>
<div>
<button>Reverse Text</button>
</div>
</form>
{ /* Render reversed text */}
<p>Reversed Text: {this.state.reversedText}</p>
</React.Fragment>
);
}
}}
No, it won’t because you can only listen to events on DOM elements. We touched on this at the beginning of the post, but React components are wrappers for DOM elements. That means we essentially have a layer that we need to pass through to listen for the event.
The way around this is to pass the event handler as a prop to the child component. Then the prop is passed down to the click event as an attribute like so:
There may be times when you want to make use of certain DOM events that are triggered when the component is mounted. Let’s see this using the resize event — we want to see the width of the window whenever it is resized.
If we create a component and try it out like we have below, then the event will not be triggered. We’ll need to add the event listener (handleResize() in this case) and the event type like we have here:
Now, the event listener will be added when the component mounts. That means our component is actively listening to the browser window and will display its width when it updates.
In summary
OK, so we covered quite a bit of ground in a very small amount of space. We learned that React does not connect directly to a DOM event, but rather Synthetic Events that are wrappers for DOM events. We dug into the process for creating event listeners so that they attach to Synthetic Events and, from there, made sure that a component will update when those events are triggered.
