Front-End of Future Past

Slides from my talk at the Big (D) Conference, 2014.

Transcript

1. F R O N T - E N D O F F U T U R E P A S T
2. #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
4. 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.
5. * 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.*
6. Difficult to see. Always in motion is the future.
7. To determine where we may be heading, let's reflect on where we've been…
8. PART 1: HOW WE GOT HERE
9. http:/ /w3.org/People/Raggett/book4/ch02.html HTML
10. http:/ /w3.org/People/Berners-Lee 1989: Tim Berners-Lee created HTML to link research docs.
11. "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
12. http:/ /en.wikipedia.org/wiki/CERN HTML was for serious work. Initially, no <img> tag.
13. Before long though, the masses got their hands it. That is when great chaos befell the Internet…
15. But, I digress. We're getting a little ahead of ourselves. ( sorry )
16. http:/ /en.wikipedia.org/wiki/Mosaic_(web_browser) Marc L. Andreessen Eric J. Bina
17. "Standardizing" on browser innovation… 1993: Marc Andreessen proposed the <IMG> tag. Mosaic later becomes Netscape, which then introduces JS, SSL, etc.
19. JavaScript http:/ /oreilly.com/pub/a/javascript/2001/04/06/js_history.html
20. "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
21. JavaScript is quirky and prototypal. Java is classy. Get it? :)
22. 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 :)
23. http:/ /en.wikipedia.org/wiki/Netscape_Navigator 1995: JavaScript debuted in Netscape Navigator 2.0.
24. 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...
25. 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.
26. With apologies to Jeremy Keith and Brad Colbow… http:/ /coding.smashingmagazine.com/2009/07/29/misunderstanding-markup-xhtml-2-comic-strip
27. 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
28. 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
29. 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)
30. 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
31. 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
32. 2005: Then there was jQuery... Heck yeah! $('p.neat').addClass('ohmy').show('slow'); JavaScript that reads like CSS code!? http:/ /jquery.com
33. $(document).ready(function() { // Many tutorials put all the code here. }); But we treated it like CSS... Um, okay... What's the big deal?
34. 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)
35. 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
36. http:/ /w3.org/Style/LieBos2e/history/Overview.html CSS
37. http:/ /intercom.co.cr/www-archives/1994-q1/0648.html 1994: Marc Andreessen, on styling... "Sorry, you're screwed."
38. 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
39. 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
40. Brief history of the first "Browser War" http:/ /en.wikipedia.org/wiki/Browser_wars
41. 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.
42. We made sites "best viewed in IE6," and Netscape lost the browser war. http:/ /en.wikipedia.org/wiki/Netscape_Navigator
43. 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
44. In 2001, IE had over 90% market share. https:/ /en.wikipedia.org/wiki/Browser_wars
45. 2004: Firefox arose from Netscape's ashes. Thus began a new Browser War, still in progress.
46. Since then, IE has been in steady decline. http:/ /en.wikipedia.org/wiki/File:Usage_share_of_web_browsers_(Source_Net_Applications).svg
47. Chrome grew, at the expense of Firefox. http:/ /en.wikipedia.org/wiki/File:Usage_share_of_alternative_web_browsers_(Source_StatCounter).svg
48. Today, Chrome has overtaken IE… What!? https:/ /en.wikipedia.org/wiki/Browser_wars Chrome = green | IE = blue | Firefox = orange
49. "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?
50. PART 2: WHERE WE'RE GOING
51. http:/ /zeldman.com/2009/07/02/xhtml-wtf
52. 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!
53. There were differing schools of thought on HTML vs. XHTML. In the field "make it work" Academic "make it valid"
54. rapid —> iteration http:/ /whatwg.org/specs/web-apps/current-work/multipage
55. Ian Hickson, benevolent HTML5 specification dictator curator… http:/ /en.wikipedia.org/wiki/Ian_Hickson
56. 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…
57. 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.
58. 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>
59. 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?
60. 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
61. HTML5 CSS3 http:/ /w3.org/html/logo What about JS? It's actually not stewarded by the W3C.
63. https:/ /github.com/voodootikigod/logo.js
64. Branding, FTW! So, what's next? 1. Get cool logo(s). [X] 2. Convince others. [X] 3. Profit! [X]
65. http:/ /apple.com/html5 Apple likes HTML5.
66. http:/ /html5rocks.com Google likes HTML5.
67. https:/ /developer.mozilla.org/en/HTML/HTML5 Mozilla likes HTML5.
68. http:/ /opera.com/docs/specs/presto25/html5 Opera likes HTML5.
69. http:/ /msdn.microsoft.com/en-us/ie/aa740476 Microsoft likes HTML5.
70. HTML5 <canvas> is gaining traction. http:/ /playbiolab.com
71. As of IE9, all browsers support <canvas>. http:/ /ie.microsoft.com/testdrive/performance/asteroidbelt/default.html
72. 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
73. 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.
74. http:/ /flickr.com/photos/azuriblue/3350954960 Canvas is like a whiteboard drawing, one layer. Note: Canvas only works in the browser via JS.
75. Quick CSS tip to fix IE's SVG rendering: svg { overflow: hidden; }
76. IE9, without the CSS fix for SVG: http:/ /host.sonspring.com/raph
77. IE9, with the CSS fix for SVG: http:/ /host.sonspring.com/raph
78. SVG in: Chrome, Firefox, Opera, & Safari. http:/ /host.sonspring.com/raph
79. 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
80. More Semantic Richness!
 New tags in HTML5 http:/ /html5doctor.com/element-index
81. <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 )
82. 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
83. https:/ /github.com/mrdoob/three.js
84. https:/ /github.com/mrdoob/three.js
87. 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
88. 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.
89. 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
90. 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.
91. Note: Don't indent unnecessarily with Sass. // This looks innocent enough: table.table-class tbody tr td a.link-class color: orange
92. 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; }
93. Haml rocks, if you're in a Ruby project. To me, Haml means never typing this again... </div></div></div>
94. Haml can generate nicely formatted HTML. I'm not an über-Ruby guy. Get your feet wet:
95. 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.
96. get.your(coffee) { return 'off my lawn'; }
97. Not everyone's a fan. Dustin Diaz, quoting Douglas Crockford:
98. http:/ /blog.meloncard.com/post/12175941935/how-one-missing-var-ruined-our-launch
99. http:/ /nodejs.org
100. 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...
101. 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
102. 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
103. Recursion -> CoffeeScript is written in… CoffeeScript!? https:/ /plus.google.com/116264189418994838408/posts/CSXeyftovTJ
104. # CoffeeScript: $('#foobar').click -> # Do stuff. // JavaScript: $('#foobar').click(function() { // Do stuff. }); CoffeeScript -to- JavaScript... LOL - That's all? http:/ /coffeescript.org
105. # 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
106. "Unless" is cool… # CoffeeScript: happy = true unless raining // JavaScript: var happy; if (!raining) { happy = true; } That's hawt. http:/ /coffeescript.org
107. 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.
108. 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?
110. 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…
111. 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.
112. This is why we can't have nice things!
113. Backbone.js — Actually pretty nice…
114. Underscore.js — Also nice…
115. But use Lo-Dash instead. It's more solid…
116. There are plenty of good MV* JS frameworks. &
121. 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
122. Trask-Industries.com (X-Men spoof site) uses Angular.js
123. Trask-Industries.com (X-Men spoof site) uses Angular.js
124. DEMO http://github.com/nathansmith/ko-table
126. 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.
127. <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>
128. <tbody data-bind="foreach: data"> <tr> <td data-th="First Name:"> <span data-bind="text: first_name || '&mdash;'"></span> </td> <td data-th="Last Name:"> <span data-bind="text: last_name || '&mdash;'"></span> </td> <td data-th="Character First Name:"> <span data-bind="text: character_first_name || '&mdash;'"></span> </td> <td data-th="Character Last Name:"> <span data-bind="text: character_last_name || '&mdash;'"></span> </td> </tr> </tbody>
129. <tbody data-bind="foreach: data"> <tr> <td data-th="First Name:"> <span data-bind="text: first_name || '&mdash;'"></span> </td> <td data-th="Last Name:"> <span data-bind="text: last_name || '&mdash;'"></span> </td> <td data-th="Character First Name:"> <span data-bind="text: character_first_name || '&mdash;'"></span> </td> <td data-th="Character Last Name:"> <span data-bind="text: character_last_name || '&mdash;'"></span> </td> </tr> </tbody>
130. // 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" }, ... ]
131. // 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; }); };
132. // 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; } }, ...
133. // 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); }); },
134. 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