Share
Embed
Front-End of Future Past
Slides from my talk at the Big (D) Conference, 2014.
Nathan Smith
September 06, 2014
Transcript
134 slides
-
F R O N T - E N D O
F F U T U R E P A S T -
#whoami ? My name is Nathan. I do UX and
JS at TandemSeven. But most people know me as "the 960 guy" (that's ok). http:/ /sonspring.com | http:/ /960.gs | http:/ /tandemseven.com -
Download the slides and follow along: http:/ /j.mp/future-past Or, talk
smack about me on Twitter: @nathansmith Note: You don't have to take notes feverishly. -
* My predictive powers are average (at best). Today, I
want to talk a bit about what the future of HTML, CSS, and JavaScript will look like.* -
To determine where we may be heading, let's reflect on
where we've been… -
"Storage of ASCII text, and display on [80x24] screens, is
in the short term sufficient, and essential. Addition of graphics would be an optional extra with very much less penetration for the moment." – Tim Berners-Lee http:/ /w3.org/History/1989/proposal.html -
Before long though, the masses got their hands it. That
is when great chaos befell the Internet… -
"Standardizing" on browser innovation… 1993: Marc Andreessen proposed the <IMG>
tag. Mosaic later becomes Netscape, which then introduces JS, SSL, etc. -
"JavaScript had to 'look like Java' only less so, be
Java's dumb kid brother or boy- hostage sidekick. Plus, I had to be done in ten days or something worse than JS would have happened." Brendan Eich created JavaScript in 10 days. http:/ /en.wikipedia.org/wiki/Brendan_Eich -
With classical inheritance, if your mother breaks her leg, you
go visit her in the hospital. With prototypal inheritance, your leg would also (instantly) break. Object Inheritance #TLDR #TLDR = "Too Long, Didn't Read" aka: Sum it up for me :) -
The word "JavaScript" was chosen to tap into the brand
name of the Java programming language. Sun Microsystems (since acquired by Oracle) owned the trademarks for both Java™ and JavaScript™, and stewards the development of Java (not JavaScript). In 1997, JavaScript (the language) was standardized by the European Computer Manufacturers Association, and it was dubbed "ECMAScript." http:/ /en.wikipedia.org/wiki/JavaScript | http:/ /en.wikipedia.org/wiki/ECMAScript Um - what? Brief history of JavaScript = Clear as mud... -
Altogether different beasts: Both potentially delicious (or dangerous), but not
actually derived from the same animal. Similar in name: JavaScript is to Java as Hamburger is to Ham. -
1996: IE 3.0 ships with "JScript" (for copyright purposes). In
1995, MS begins to ship Internet Explorer with Windows 95. For Netscape, it was the beginning of the end. Microsoft also reverse- engineered JavaScript, as "JScript." http:/ /en.wikipedia.org/wiki/Internet_Explorer -
Created at Netscape "ECMA TC39" Committee responsible for evolving the
language Who is responsible for JavaScript now? {to name a few... http:/ /en.wikipedia.org/wiki/JavaScript | http:/ /en.wikipedia.org/wiki/ECMAScript TM of Oracle (via Sun) Defines the ECMAScript DOM API -
JS browser support is pretty good. JavaScript (ECMAScript 3.0) is
supported fairly consistently in all major browsers. ⾨ This is Wikipedia's compatibility table* for: – Trident (IE) – Gecko (Firefox) – WebKit (Chrome & Safari) – Presto (Opera) * I realize those are rendering engines, not specific JS engines (which change names more rapidly). It's how Wikipedia lists 'em. http:/ /en.wikipedia.org/wiki/Comparison_of_layout_engines_(ECMAScript) -
JavaScript is that friend in high school you secretly had
a crush on, but was always hanging out with the sleazy boyfriend, DOM. Nice > What a punk! http:/ /w3.org/DOM -
The Document Object Model (DOM) is what allows JS to
interact with XML/HTML, ala: window.document.getElementsByTagName('*'); The DOM is stewarded by a different group than governs JS. Browsers also implement it slightly differently. For instance, IE historically didn't treat whitespace as text nodes, even though the official DOM spec says it should. http:/ /w3.org/DOM -
2005: Then there was jQuery... Heck yeah! $('p.neat').addClass('ohmy').show('slow'); JavaScript that
reads like CSS code!? http:/ /jquery.com -
$(document).ready(function() { // Many tutorials put all the code here.
}); But we treated it like CSS... Um, okay... What's the big deal? -
Loosely translated: "You'd better lose yourself in the music, the
moment. You own it, you better never let it go... You only get one shot, do not miss your chance to [load]." (with apologies to Eminem) window.onload http:/ /en.wikipedia.org/wiki/8_Mile_(film) -
Really, it's not the fault of jQuery, nor of any
other capable JavaScript libraries: Dojo, MooTools, Prototype, YUI, etc. As is almost always the case with code, the problem is actually a human one. We've become complacent and forgotten (or never learned) how JavaScript works. Beware the danger of abstraction & illiteracy. http:/ /ejohn.org/blog/javascript-language-abstractions -
1994: Håkon Wium Lie proposed CSS. Initially it was called
CHSS, ("Cascading HTML Style Sheets"), but was later shortened to simply CSS. 1996: The first browser to support CSS was Internet Explorer 3.0, followed closely by Netscape Navigator version 4.0. http:/ /people.opera.com/howcome/2006/phd -
CHSS had 100 levels of !important h1.font.size = 24px 12%
/* 12% importance */ h1.font.size = 16px 100% /* 100% importance */ h1.font.size = 12px 89% /* 89% importance */ h1 {font-size: 24px;} /* Applies globally (weak) */ h1.x {font-size: 16px;} /* More specific */ h1 {font-size: 12px;} /* Beats line 1, not line 2 */ Thankfully, CHSS was never implemented: CSS has "last-in" wins, with specificity override. http:/ /www.w3.org/People/howcome/p/cascade.html -
1998: "Web Standards" advocacy begins. http:/ /w3.org/Style/LieBos2e/history/Overview.html The box model
isn't fixed until October 2006, in IE7 debut. -
We made sites "best viewed in IE6," and Netscape lost
the browser war. http:/ /en.wikipedia.org/wiki/Netscape_Navigator -
The "dark side" = developing for 1 browser. Oh, what...
Like you never photoshopped you vs. yourself with a lightsaber? C'mon! http:/ /flickr.com/photos/nathansmith/32757176 -
2004: Firefox arose from Netscape's ashes. Thus began a new
Browser War, still in progress. -
Today, Chrome has overtaken IE… What!? https:/ /en.wikipedia.org/wiki/Browser_wars Chrome =
green | IE = blue | Firefox = orange -
"Those who cannot remember the past are condemned to repeat
it." – George Santayana http:/ /en.wikipedia.org/wiki/George_Santayana So your "pure CSS" demo only works in WebKit? -
Realistically speaking, XHTML 2.0 would've been a very tough sell,
from a practical standpoint. It was extremely rigid. A page simply would not render, if it wasn't perfectly formed from an XML structural standpoint. Oh, and it would've done away with <img>. Oh noes! -
There were differing schools of thought on HTML vs. XHTML.
In the field "make it work" Academic "make it valid" -
Technically, HTML5 is now a misnomer. (But, it's too late.
Marketing already got ahold of it.) HTML is to be considered an ever-evolving "living standard." The WHATWG — Web Hypertext Application Technology Working Group — will continue to work on HTML, and the W3C will snapshot feature sets periodically, and tag them with incremental version numbers: HTML5 (and presumably, HTML6, etc). Thus, no more version numbers in DOCTYPE… -
HTML5 "killed" XHTML. Good riddance, I say. <!DOCTYPE html> <!DOCTYPE
html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/ DTD/xhtml1-transitional.dtd"> <— Much easier to remember! As it turns out, <!DOCTYPE html> is the minimum number of characters to trigger "standards" mode in older IE versions. -
You don't have to type extra junk anymore. <script> alert('This
is some inline JS.'); </script> <script type="text/javascript" charset="utf-8"> /* <![CDATA[ */ alert('Are you kidding me?'); /* ]]> */ </script> -
Have you ever used a non-CSS stylesheet? <link rel="stylesheet" href="/path/file.css"
/> <link rel="stylesheet" href="/path/file.css" type="text/css" media="all" /> Note: Media queries are best defined via @media in a single stylesheet anyway, to avoid unnecessary HTTP requests. No? Me neither, so why "type" it? -
You can't do that! What about the validator? Relax. It's
valid. HTML5 is no longer a sub- set of SGML like all versions of HTML before it. Nor is it a spin on XML, like XHTML 1.x was. HTML is just HTML now. This is how browsers always treated (X)HTML anyway. I know what you XHTML purists are thinking: ( phew! ) http:/ /en.wikipedia.org/wiki/HTML -
HTML5 CSS3 http:/ /w3.org/html/logo What about JS? It's actually not
stewarded by the W3C. -
Branding, FTW! So, what's next? 1. Get cool logo(s). [X]
2. Convince others. [X] 3. Profit! [X] -
2011: IE9 supports SVG (proposed in 1999). Examples of Dojo.js
charts To be fair, IE renders SVG very fast, via hardware acceleration. Konqueror (WebKit predecessor), was the first browser to have built-in SVG support, in 2004. http:/ /user.sitepen.com/~dwalsh/dojo-charting.html -
http:/ /flickr.com/photos/124330160/2127121118 And, each layer can respond to mouse events!
SVG is like construction paper. Layers retain distinct shapes. SVG files work fine without JS. -
http:/ /flickr.com/photos/azuriblue/3350954960 Canvas is like a whiteboard drawing, one layer.
Note: Canvas only works in the browser via JS. -
Quick CSS tip to fix IE's SVG rendering: svg {
overflow: hidden; } -
I won't cover <svg> or <canvas> in detail during this
talk, but two helpful JavaScript libraries are worth noting... Raphaël (SVG) – Dmitry Baranovskiy Processing JS (Canvas) – John Resig http:/ /raphaeljs.com http:/ /processingjs.org -
<header role="banner"> <footer role="contentinfo"> <aside> <nav role="navigation"> <main role="main"> <article>
<section> <hgroup> <section> <dialog> <p> etc... <figure> <img /> <details> <summary> <p> etc... <figcaption> Mixes well with ARIA roles ( Accessible Rich Internet Applications ) -
The dawn of 3D browser graphics. aka: Can I play
Farmville in IMAX!? WebGL is going to be HUGE ( eventually ) http:/ /en.wikipedia.org/wiki/WebGL -
Thoughts on Metaprogramming… Haml => HTML Sass/Compass => CSS CoffeeScript
=> JS Note: Yes, I'm sure there are many, many more examples in the wild. But my time is finite, and this is what I'm familiar with. http:/ /haml-lang.com | http:/ /sass-lang.com | http:/ /compass-style.org | http:/ /coffeescript.org -
Personal note: After giving 'em a try, I am on
the fence regarding meta languages that treat whitespace as significant. Usual disclaimers apply: "It depends." Your mileage may vary. No purchase necessary. Machine-wash only. Enter at own risk. Batteries not included. Void where prohibited, etc. -
Example of Sass (with Compass). @import compass/css3 // CSS3 helpers,
etc. // I write this Sass code: .foobar +background-image(linear-gradient(#fff, #ccc)) +border-top-right-radius(4px) +box-shadow(rgba(#000, 0.5) 0 2px 5px 0) &:hover text-decoration: underline -
Example of Sass (with Compass). /* This CSS is generated...
*/ .foobar { background-image: -webkit-gradient(linear, 50% 0%, 50% 100%, color-stop(0%, #ffffff), color-stop(100%, #cccccc)); background-image: -webkit-linear-gradient(#ffffff, #cccccc); background-image: -moz-linear-gradient(#ffffff, #cccccc); background-image: -o-linear-gradient(#ffffff, #cccccc); background-image: -ms-linear-gradient(#ffffff, #cccccc); background-image: linear-gradient(#ffffff, #cccccc); -moz-border-radius-topright: 4px; -webkit-border-top-right-radius: 4px; -o-border-top-right-radius: 4px; -ms-border-top-right-radius: 4px; -khtml-border-top-right-radius: 4px; border-top-right-radius: 4px; -moz-box-shadow: rgba(0, 0, 0, 0.5) 0 2px 5px 0; -webkit-box-shadow: rgba(0, 0, 0, 0.5) 0 2px 5px 0; -o-box-shadow: rgba(0, 0, 0, 0.5) 0 2px 5px 0; box-shadow: rgba(0, 0, 0, 0.5) 0 2px 5px 0; } .foobar:hover { text-decoration: underline; } BLOAT? Not really. You'd have to type all this to make it work cross-browser anyway. -
Note: Don't indent unnecessarily with Sass. // This looks innocent
enough: table.table-class tbody tr td a.link-class color: orange -
Note: Don't indent unnecessarily with Sass. /* But it generates
this... */ table.table-class tbody tr td a.link-class { color: orange; } /* When you probably just meant... */ .table-class .link-class { color: orange; } -
Haml rocks, if you're in a Ruby project. To me,
Haml means never typing this again... </div></div></div> -
Haml can generate nicely formatted HTML. I'm not an über-Ruby
guy. Get your feet wet: -
Okay, so abstraction for HTML and CSS is fine. It
saves keystrokes, and those languages are pretty #yawn anyway, right? I mean, they're declarative, so of course we should automate them. But do not touch my JavaScript. -
Can you spot the bug? var joe = 1 sue
= 2, bob = 3; var joe = 1, sue = 2, bob = 3; Bugs like this can be laborious to find, which is why I write... -
It's okay to be obvious. var joe = 1; var
sue = 2; var bob = 3; Minifiers like Uglify JS can handle removing repeated var keywords. Put that burden on automation, not yourself. https:/ /github.com/mishoo/UglifyJS | http:/ /marijnhaverbeke.nl/uglifyjs -
Just tell JSLint… "Relax, it's okay." JSLint (and JSHint) is
a great tool for checking for JS errors. But validation is a means to an end, not a goal in itself. If you want to ensure JSLint compliance, there's CoffeeScript... http:/ /jslint.com | http:/ /jshint.com -
# CoffeeScript: $('#foobar').click -> # Do stuff. // JavaScript: $('#foobar').click(function()
{ // Do stuff. }); CoffeeScript -to- JavaScript... LOL - That's all? http:/ /coffeescript.org -
# CoffeeScript: is_between = 1 < my_value < 10 //
JavaScript: var is_between; is_between = (1 < my_value && my_value < 10); CoffeeScript = Syntactic Sugar... Oh, I see. http:/ /coffeescript.org -
"Unless" is cool… # CoffeeScript: happy = true unless raining
// JavaScript: var happy; if (!raining) { happy = true; } That's hawt. http:/ /coffeescript.org -
Despite its positive aspects, I am not a fan of
CoffeeScript. Why? Because it's more than just syntactic sugar. It's dangerous artificial sweetener, so to speak. -
http:/ /coffeescript.org obj = foo: 'bar' for i in obj
console.log obj[i] var i, obj, _i, _len; obj = { foo: 'bar' }; for (_i = 0, _len = obj.length; _i < _len; _i++) { i = obj[_i]; console.log(obj[i]); } Want to iterate through an object in CoffeeScript? Sure thing, just... WTF? -
http:/ /www.2ality.com/2013/06/iterators-generators.html let arr = ['foo', 'bar', 'baz']; for (let
[index, element] of arr.entries()) { console.log(index + '. ' + element); } // Yields... // // 1. foo // 2. bar // 3. baz "for of" loops in the future… -
http:/ /coffeescript.org obj = foo: 'bar' for i of obj
console.log obj[i] var i, obj; obj = { foo: 'bar' }; for (i in obj) { console.log(obj[i]); } You would have to actually type this in CoffeeScript. Then the resulting JavaScript output is a "for in" loop. -
Key tenets of most MV* JS frameworks: — Model, View,
Controller (MVC) or… — Model, View, View Model (MVVM) — Two-way data binding… If user interacts with page, you can reflect these changes in your data — Declarative UI: in markup, not in JS — Observables: If data changes, UI updates -
Below a certain width, the layout switches to a “mobile”
view. The table rows & cells are display:block, and text from each <th> is inserted as a label, preceding the data. -
<thead> <tr> <th scope="col" data-key="first_name"> <div class="cell"> <a href="#">First Name</a>
</div> </th> <th scope="col" data-key="last_name"> <div class="cell"> <a href="#">Last Name</a> </div> </th> <th scope="col" data-key="character_first_name"> <div class="cell"> <a href="#">Character First Name</a> </div> </th> <th scope="col" data-key="character_last_name"> <div class="cell"> <a href="#">Character Last Name</a> </div> </th> </tr> </thead> -
<tbody data-bind="foreach: data"> <tr> <td data-th="First Name:"> <span data-bind="text: first_name
|| '—'"></span> </td> <td data-th="Last Name:"> <span data-bind="text: last_name || '—'"></span> </td> <td data-th="Character First Name:"> <span data-bind="text: character_first_name || '—'"></span> </td> <td data-th="Character Last Name:"> <span data-bind="text: character_last_name || '—'"></span> </td> </tr> </tbody> -
<tbody data-bind="foreach: data"> <tr> <td data-th="First Name:"> <span data-bind="text: first_name
|| '—'"></span> </td> <td data-th="Last Name:"> <span data-bind="text: last_name || '—'"></span> </td> <td data-th="Character First Name:"> <span data-bind="text: character_first_name || '—'"></span> </td> <td data-th="Character Last Name:"> <span data-bind="text: character_last_name || '—'"></span> </td> </tr> </tbody> -
// In a real app, this data would potentially be
dynamic. // But for the purposes of this demo, is hard-coded here. [ { "first_name": "Amy", "last_name": "Poehler", "character_first_name": "Leslie", "character_last_name": "Knope" }, { "first_name": "Nick", "last_name": "Offerman", "character_first_name": "Ron", "character_last_name": "Swanson" }, { "first_name": "Aziz", "last_name": "Ansari", "character_first_name": "Tom", "character_last_name": "Haverford" }, ... ] -
// Extend KO array, to make it sortable ko.observableArray.fn.sort_by =
function(key, reverse) { var self = this; self.sort(function(a, b) { var a_key = String(a[key]); var b_key = String(b[key]); var n, val; if (reverse) { n = a_key - b_key; val = !isNaN(n) ? n : b_key.localeCompare(a_key); } else { n = b_key - a_key; val = !isNaN(n) ? n : a_key.localeCompare(b_key); } return val; }); }; -
// APP.models models: { // APP.models.table_view_model table_view_model: function() { var
self = this; // This data comes from "/json/data.js" APP.data = APP.data || ko.observableArray(DATA_JSON); self.data = APP.data; } }, ... -
// APP.init.sort_by sort_by: function(key) { var event = 'click.sort_by'; var
str = '.table-data th[data-key] a'; body.off(event).on(event, str, function(ev) { var el = $(this); var th = el.closest('th'); var th_other = th.siblings('th'); var key = th.attr('data-key'); var sort = th.attr('data-sort'); var asc = 'ascending'; var desc = 'descending'; var dir = asc; if (!sort || sort === asc) { dir = desc; } var reverse = dir !== asc; th.addClass(on).attr('data-sort', dir); th_other.removeClass(on).removeAttr('data-sort'); APP.data.sort_by(key, reverse); }); }, -
Get the slides… http:/ /j.mp/future-past Keep in touch… @nathansmith Questions?
— I may (not) know the answer! http:/ /corsariomarcio.deviantart.com/art/X-MEN-DAYS-OF-FUTURE-PAST-poster-455872133