jQuery and the W3C

29
jQuery and Standards John Resig http://ejohn.org / - http://twitter.com/jeresig

Transcript of jQuery and the W3C

jQuery and StandardsJohn Resig

http://ejohn.org/ - http://twitter.com/jeresig

jQuery✦ A JavaScript library designed to hide

painful cross-browser compatibility issues while presenting a solid, usable, API.

Simple API✦ $(“div > span”).addClass(“foo”);

✦ “Find some elements”✦ “Do something with them”

✦ Makes complex manipulation of web pages simple

Space✦ Highly competitive space✦ Released Jan. 2006 - already a dominant

player: Prototype JavaScript Library✦ (Bundled with Ruby on Rails, had some

nice coattail growth.)✦ Other libraries: Dojo, Yahoo UI,

MooTools

Specifications✦ A few specifications that matter to us:

✦ DOM✦ HTML✦ CSS✦ ECMAScript

✦ DOM more than anything else.

Concerns✦ Performance.✦ Performance.✦ Performance.✦ Usability.✦ Any standards/browser addition that gives

us performance benefits we’ll leap on.

New Standards We’ve Used✦ Selectors API

✦ querySelectorAll✦ Selectors API 2

✦ matchesSelector✦ Animation Timing

✦ requestAnimationFrame✦ ECMAScript

✦ bind

Selectors API✦ A bit of a failure✦ Didn’t listen to the needs of libraries✦ Missed a number of important features/

bug fixes:✦ Contextual searching is messed up✦ Error reporting is non-existent✦ Implementations are inconsistent

✦ But it’s very fast, so we use it.

Matches Selector✦ Selectors API 2 gave us matchesSelector✦ We petitioned browsers to implement this✦ They did, then it became a standard✦ Makes our event delegation much faster

Smooth Animation✦ requestAnimationFrame was created✦ Scales animations based upon load✦ Unfortunately this broke user expectations (expected certain frame rates)

✦ We just backed it out, will have to try again later

Needs

HTML string -> DOM support✦ No good way to do this now✦ Have to create a DOM element and use

innerHTML✦ Clunky and quite slow✦ We want:

✦ someMethod(“<b>stuff</b>”) ->✦ [ <b> ]

Access to event callbacks✦ We want to be able to remove individual

callbacks✦ We want to be able to clone callbacks✦ We want to be able to trigger specific

callbacks✦ All of this requires access to callbacks

An event for when stylesheets load

✦ Right now we have an event for DOM loaded

✦ And an event for window loaded✦ But no event for when all the stylesheets

load (important for looking at computed styles)

Will an element fire an event?✦ For example - if I have a <form> element I

want to be able to ask it:✦ “Will you ever, natively, trigger a submit

event?” (true)✦ If I ask a <div> if it will trigger a submit

event, it will return false.

Unique ID for DOM nodes✦ We have to manage callbacks and data that

we attach to DOM nodes✦ To do this we have to assign the nodes a

unique ID✦ It’d be much better to have a property that

took care of this for us

“Late Events”✦ There is no way to ask the browser:

✦ “Did an event [foo] already fire on this element?”

✦ For example:✦ Did the load event already fire on

window?✦ Did the submit event already fire on this

form?✦ etc.

Fast DOM mutation events✦ I know this is being worked on right now (yay!)

✦ A way to have fast DOM mutation events would be awesome

✦ It could allow for some really slick restructuring of applications

✦ And make it easier for us to possibly do caching

mouseenter/mouseleave✦ Internet Explorer provides these events✦ They’re terribly useful (make it so that you

don’t have to deal with event bubbling weirdness)

✦ Should be in browsers✦ Need to verify DOM 3 Events spec

getComputedStyle✦ A complete mess right now✦ There is no consensus over what results

should be returned and when✦ There needs to be something declared

here - probably a joint venture between the CSS and DOM working groups.

✦ Test suite for getComputedStyle

isCSSAuto✦ There is no way of determining if a CSS

property is currently set to “auto”✦ This should be resolved, makes it much

easier to do some types of animations

A way to sanely toggle visibility✦ If we’re given an element that is display:

none and we want to make it visible (display: block, perhaps)

✦ It is very hard to determine what the right “visible” style should be

✦ Especially if someone does:✦ div { display: none; }

✦ Hint: It involves nasty use of iframes

contains() method✦ We have compareDocumentPosition✦ This is OK but contains() is very easy to

use (in IE)✦ Easy enough to implement, should be a

standard

Better way of sorting nodes✦ We have to use

compareDocumentPosition now✦ This is very very slow✦ A numerical index property on nodes

would be very useful (like in IE)

Is a node in an XML document✦ A number of behaviors change when you’re

in an XML document✦ (IDs no longer resolve, some properties

may not exist - like innerHTML, etc.)✦ A way to determine if we’re working

against an XML document would help

Support Tests