FFconf 2017 talks in 280 characters (or close enough)
James Kyle on web dev tooling
Software engineering is traditionally taught starting with the fundamental concepts: first, learn what ifs and linked lists are, then create some simple app.
But would it be better to start learning (and building!) with create-react-app instead? Yes, it would.
Once somebody starts learning, let’s not to put them off with confusing competing standards like webpack, browserify and lasso.
James is on it.
Bruce Lawson on CSS and humans
Everything we do as developers is ultimately for humans, and these users don’t care what library we used to build the website. They care if app does what they need.
Tools and frameworks are useful, though — let’s avoid the hacky ones (cough CSS-in-JS cough) and use those built upon standards instead.
Something like Bruce’s own stylable, perhaps?
Monica Dinculescu on how to make web standards
The process for new standards has learned a lot since the days of Application Cache and Web Intents. Low level APIs like Service Workers are now recognised as the better option, with new features following their lead.
Fancy creating a web standard yourself?
Gone are the days when those were discussed over obscure email distribution lists — we can totally do that on the interwebs now!
What about when a new feature comes out but not all browsers support it yet?
polyfill.io is here to help.
Addy Osmani on making websites fast
Network calls and CPU tasks such as script parsing/execution or image decoding are the main reasons websites are slow.
The PRPL pattern helps to minimise their impact: load immediately only what you really need and wait for browser idle time to fetch the rest.
Some tools that help speed things up:
Once happy with site speed, set a performance budget to ensure it won’t slow down over time.
Blaine on passwords
Frustration at length and mixed character password requirements are common, but what if you wanted to use a (far safer!) combination of arabic/mandarin characters and a 🐶?
As it turns out, that won’t work most of the time.
Short session durations don’t help either — users then have to type their pass all the time, which makes them prime phishing targets.
What can you do then? Use centralised authentication, like Gmail/Facebook auth.
@katie_fenn on garbage collection
JS developers do not really have to think about garbage collection.
Or do they?
Any variable referencing data means that that data is being used. What if you leave a global variable around that’s not being used anymore?
The garbage collector won’t be able to tell if it can be deleted, so it’ll leave it alone.
The solution? Use closures instead, and make sure not to create and delete lots of data very quickly — GC hates that.
Jenn Schiffer on pixel art, ethics and user content
Remember when we could see nicely formatted source code of pretty much any html/JS/CSS out there?
That was a super useful way to learn how things work, but minification took that away. Source maps are nice, but not quite as good.
Joe Hart on er, hmm, web games comedy?
We had a ~100 people strong flappy birds multiplayer game.
We clapped, it flapped on the giant screen.
Best. Talk. Ever.
And that’s all, folks.
Thank you Julie Sharp and Remy Sharp for organising an amazing conf!
