Explore My Notes

Emoji before emoji | ilovetypography

Rebus writing substitutes pictures or symbols for words, but not in the same way that pictograms do. With pictograms, a picture of, for example, a bee simply represents the insect. But in rebus writing, a picture of a bee is used to substitute for the letter b or its sound... a picture of an eye represents the letter i, and so on.

In other words, text speech or a variant on modern emoji speech. Cool/❄ 😄

A tale of two buttons | James Nash

You have buttons on your website. These can be on a light background, or a dark background, and have active and non-active states, but should be highly visible in any variation. You might think that means you need four CSS rules, but James does a great job of showing some tricks that mean you can achieve all of that with just one (plus a couple of custom properties). Simple tools, like inherit, are used to great effect. I mean I rarely use that attribute, yet James make it look like an indispensable superpower! 😅 Very clever indeed and a strong argument for why using the cascade to your advantage is often more beneficial than completely destructuring it with systems like BEM.

Remember, every time you set a property in CSS you are in fact overriding something (even if it’s just the default user agent styles). In other words, CSS code is mostly expressing exceptions to a default design. The more you do so, the more your UI components become sealed-off blocks of UI that don’t necessarily blend into or adapt to their surroundings.

Accessible interactions | adactio

So you have a non-page-based navigation: tabs, carousels, accordions, multi-level menus, route changes... we deal with these a lot in modern frontends. Is your "trigger" (the thing you press to make the navigation happen) a link or a button? I tend to side with link (it's taking you to a part of the document, just a hidden one right now). Jeremy has a slightly more considered approach that thinks less about semantics (without ignoring them) and more about actual user experience:

  • Is the target of the trigger close by, or directly adjacent to, the trigger in the DOM? Use a button, because you don't need to worry about a focus change.
  • Otherwise, use a link with a fragment identifier (y'know, href='#target').

As he points out, that means you may end up using a link even where a button is sort of semantically cleaner, such as a login modal. In my head, a modal is always a button: it doesn't take me to a part of the document or another page, it opens a new form entirely (Jeremy sees this differently anyway, and I can understand his reasoning). But using a link means your focus state gets switched to the modal immediately, no extra code required, plus if JS fails the link may still work.

This mental model only gets you so far, but Jeremy also highlights some useful ARIA roles that can be sprinkled on top to further increase the accessibility. I'll try to keep it all in mind 😄

📆 10 Nov 2020  | 🔗

  • Inclusion
  • a11y
  • accessibility
  • button
  • link
  • modal
  • carousel
  • ARIA
  • focus 

US election through the lens of Africa | gathara

A fascinating Twitter thread detailing the 2020 US elections in the style that Western journalists report on African political stories. Really interesting to see how much the language used changes the picture that the text conveys, even with the occasional humorous bit thrown in; also slightly terrifying to see just how screwed up the US is when viewed through this lens. Far too long to save the entire thing here, but I wanted to grab a few (dozen) of the better tweets. ThreadReader link for posterity as well.

Polls are set to open in 48 hours across the US as the authoritarian regime of Donald Trump attempts to consolidate its hold over the troubled, oil-rich, nuclear-armed, north American nation. Analysts are sceptical the election will end months of political violence.
Viral video of barricades being erected around the presidential palace -known locally as the White House- in the US capital, Washington DC, has raised tensions as rumors spread that incumbent despot, Donald Trump, will not give up power if he loses tomorrow's election.
Report by African experts on US politics calls judicial independence the "weak link" in the covid-blighted nation's democracy, warning of the potential for violence if judges hand-picked by the aging autocrat, Donald Trump, determine the winner of tomorrow's election.
Aid agencies launch "Adopt-An-American" campaign which lets people virtually adopt American kids for a small donation. A website will allow donors to track children's progress, visit with US families and, as a bonus, tour US homes for an authentic American experience.

👆 This has always felt like an immensly dehumanising approach but it's still super weird to see it in this context.

Analysts warn that the fear and desperation created by the deepening US humanitarian crisis, coupled with one of the most tribalized election campaigns in its history, may lift the lid off dangerous ancient ethnic hatreds and grievances that have festered for centuries.
Tribal violence rears its ugly head in strife-torn US as clashes break out between supporters of populist dictator Donald Trump and those of opposition leader, Joe Biden, waving Black Lives Matter flags outside the fortified Green Zone in the capital, Washington DC.
Analysts warn that armed rightwing paramilitary groups, which largely support the autocrat, Donald Trump, may take advantage of his declaration of victory and allegations of vote rigging to expand their activities in the US, one of the world's top 10 banana exporters.

I get the joke here about banana republics, but also the off-hand "factoid about the country we all care about for a week or so" is so on point it's brilliant 😂

Despised by many, the Trumps have monopolized government appointments, corruptly using them to build a global business empire and enrich their friends and associates, many of whom are known criminals, while ignoring the deaths of hundreds of thousands of Americans.
In chilling scenes reminiscent of the infamous 2011 "rivers of blood" speech by Saif al-Islam Gaddafi, son of then embattled Libyan dictator, Muammar Gaddafi, Donald Trump Jr, son of embattled US dictator, Donald Trump, has vowed "total war" to keep his father in power.
Having systematically demolished public confidence in the independence and integrity of all other arms of government, US dictator Donald Trump now sets his sights on the country's courts where he will challenge his electoral loss before his hand-picked judges.
The former UK colony, with breathtaking natural beauty and richly endowed with natural resources and friendly people, but also riven by deep ethnic divisions, has suffered a series of political, humanitarian and economic crises which have killed nearly 250,000 people.
While many African observers are aghast at the conduct of the US presidential election, others note that by American standards, it was a relatively good election, in a country where gerrymandering and voter suppression are not considered reasons to invalidate elections.


📆 10 Nov 2020  | 🔗

  • Anthropocenic View
  • politics
  • America
  • USA
  • election
  • Africa
  • 2020
  • journalism
  • Twitter
  • parody
  • Donald Trump 

Practical Jamstack: intro to 11ty | James Steinbach

I "missed" the meetup (it was 2am local time 😂) but glad to catch up, a really nice introduction to 11ty/Eleventy as a build tool. The more I hear about the flexibility that 11ty gives you in terms of templating and extensibility (and the more I get fed up with JS package size coming out of some of my current solutions) the more I want to dive in.

Saving forms | adactio

A very clever (and well explained) method for helping users complete forms. Particularly with modern pop-up forms, I'd say at least once a month I'm part way through typing something and accidentally shift my mouse somewhere that causes the form to close, losing everything I had written 😡 Jeremy's trick of storing the current form (well, textarea, but it should be easy to expand) content in localStorage is a great enhancement that would solve those frustrations. Here's the gist (see what I need 😄):

  1. Use an eventListener to track when a user starts typing into a form/textarea;
  2. When that fires, set off a new eventListener to track the browser's unload event using beforeUnload (Jeremy explains why that method well);
  3. If that second listener fires, save the content of the form into localStorage against a key for the page URL/slug (you can use JSON to store multiple values in one key/value pair);
  4. If the form is submitted without issue or firing the eventListener, clear the localStorage key;
  5. Now, with that setup, you can add some initial logic that checks localStorage on page load for a key that matches the current URL and, if found, parse the JSON and populate the form.

Et voila! Neat 👍

📆 23 Oct 2020  | 🔗

  • Frontend
  • User Experience
  • LocalStorage
  • form
  • textarea
  • EventListener
  • example
  • save
  • resilient design
  • progressive enhancement
  • HTML
  • UX 

We like what we choose | Seth Godin

We like what we choose. Not the other way around. [...] If you had grown up somewhere else or some time else, there’s little doubt that you’d prefer something else. The things we think we need are simply the things we’re used to.

Why do browsers render "chuck norris" as a colour? | @foone | Twitter

Man, what a fascinating thread on the bizarre reasons that browsers will parse words into colours. To sum it up:

  • Any word will be parsed into a hex code in order for things like color: red to work with backwards compatibility in CSS;
  • The way that conversion works is to change any non-hexadecimal letter into a 0, i.e. "chuck norris" becomes "C00C0 000000";
  • But that leaves codes that are too long for the hex code format, so the next step is to shrink the string down to the maximum six characters;
  • However, rather than just shortening to six characters, first the parser increases the number of characters until the total is divisible by three, then it splits the string into thirds, then concatenates each third into two characters and stitches it all back together again 🤯
    • Why tho? 🤷‍♂️
  • So you end up in a situation where "chuck norris" becomes a vivid red, whilst "chuck norr" is a slightly ill shade of yellow...
  • And all of this matters because of hidden text in email spam (yeah, wow); if colours parse out differently, then spam is harder to identify, so the whole weird algorithm is an offical W3C standard. Brilliant 😂

The markup de-crapulator | Ian Lloyd

A humorous but useful tool that strips out all the mess that things like Styled Components and React make out of HTML and lets you properly view the source of a page. (and yeah, I use those same tools and processes, and yeah it still irks me that I'm part of the problem😁)

💩 ➡ 🥇

Three ways to autofocus in React (that almost always work) | Daniel Johnson

Whilst accessibility means you largely want to steer clear of autofocus on web forms, sometimes it can be beneficial. In HTML, we can now reach for the handy autofocus attribute, but as I recently learnt that doesn't always do what you expect when dealing with a shadow DOM (in this case, React's equivalent). Daniel's article was by far the best at both describing the issue and offering some solutions, though even here it only works most of the time (Firefox does something which I feel is logical but unfortunately not consistent with either the W3C spec or what most people would want to happen).

So here are the (confusingly more than three) techniques:

  • The autofocus attribute (😉) for input elements that are always present (elements loaded later have some weird cross-browser differences);
  • The focus() DOM API that allows you to select an element using JavaScript and force focus on it i.e. element.focus(), which works almost always but can be tricky to target correctly (and reliant on client-side JavaScript);
    • FYI remember that the opposite of focus() is blur(), and both can be handily combo'd with select() for some even better UX!
  • The React autoFocus prop, which effectively mimics the HTML autofocus attribute but only triggers it when a component mounts, helping with the cross-browser issues but meaning that if you mount before visually showing an element it can steal focus incorrectly;
  • The useRef hook, which is effectively just using the focus() API within React's own life cycle;
  • The useCallback hook, which does the same thing as useRef only with more control. This appears to be the most universal option in React right now.

Firefox and SVG fill for CSS | Matt Crouch

I ran into a bug recently where some SVGs on a website were seemingly ignoring a fill request. A quick check in dev tools showed that the CSS was being applied so... what gives? Turns out Firefox has implemented an overly literal interpretation of part of the (I believe) SVG spec, causing it to add additional child <svg> elements in certain circumstances (which are hidden from dev tools by virtue of reasons). Matt does a better job of explaining the why, but the fix is to do something like this:

.class-name use > svg {
   fill: #FFFFFF;

Or, based on our existing patches when I looked that already including overrides for g > svg and some other variations, perhaps simply:

svg.class-name * {
   fill: #FFFFFF;

Made By Me, But Made Possible By:


Build: Gatsby

Deployment: GitHub

Hosting: Netlify

Connect With Me:

Twitter Twitter

Instagram Instragram

500px 500px

GitHub GitHub

Keep Up To Date:

All Posts RSS feed.

Articles RSS feed.

Journal RSS feed.

Notes RSS feed.