50
JamesDBloom MOBILE WEB BEST PRACTICES http://about.me/jamesdbloom http://blog.jamesdbloom.com

Mobile Web Best Practices

Embed Size (px)

DESCRIPTION

The key to a successful mobile site is high performance and reliability across a wide range of device capabilities and network latencies. However, the mobile web is a hostile environment with support for HTML5, JavaScript and CSS varying widely across browsers and devices. This talk will explain best practices to build high performance mobile sites that work across a wide range of devices and capabilities. The focus will be on lessons learnt at Betfair while rewriting the entire mobile web stack and how we used techniques to maximise performance and reliability. After discussing the problems faced in mobile the talk will explain how adaptive techniques can be used to provide progressive enhancement. This will be followed by an explanation of why and where performance bottlenecks occur and how these can be solved.

Citation preview

Page 1: Mobile Web Best Practices

JamesDBloom

MOBILE WEB BEST PRACTICES

http://about.me/jamesdbloom http://blog.jamesdbloom.com

Page 2: Mobile Web Best Practices

PERCENTAGE OF MOBILE WEB

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%

http://www.tecmark.co.uk/wp-content/uploads/2011/08/Mobile-and-UK-Web-Traffic-August-2011.pdf

Page 3: Mobile Web Best Practices

¡  The Problem §  What’s different and hard about mobile

¡  The Solution §  Adaptive Techniques

§  Mobile First §  Avoid JavaScript §  Device Detection §  Adaptive Design §  Progressive Enhancement

§  High Performance §  Network

§  Less Request §  Less Bytes §  Bandwidth Efficiency §  Less Latency

§  Software §  Faster Page Rendering §  Faster Interaction

§  User §  Faster Interaction §  Improve Perception

¡  Tools §  Debugging §  Performance §  Emulators

WHAT I’M TALKING ABOUT

Page 4: Mobile Web Best Practices

¡  Network §  Different types of network:

§  GSM, GPRS, WCDMA, UMTS, EDGE, HSPA, UMTS, LTE, WiMAX §  High latencies

§  300ms to 1000ms §  High packet loss

§  From 0.1% to >20% §  Each time triggers TCP Slow Start again

§  Radio Resource Control (RRC) §  Adds 1.5s to 2s to initiate connection

§  Total Connection Time §  RRC + DNS + TCP 3-way handshake + TCP slow start + HTTP Request §  SSL adds even more

§  Bandwidth §  3G

§  Download: ~1Mbps Upload: ~0.5 Mbps §  Broadband

§  Download: ~9Mbps Upload: ~2Mbps

WHAT’S MOBILE – THE PROBLEM

Page 5: Mobile Web Best Practices

0

1000

2000

3000

4000

5000

6000

7000 SunSpider JS Benchmark (in ms) for some of the devices released in the last year

¡  Device §  Rapidly changing, widely diverse hardware §  “2500 different devices on Facebook per month” - Tobie Langel

WHAT’S MOBILE – THE PROBLEM

http://www.anandtech.com/show/5584/htcs-new-strategy-the-htc-one

Page 6: Mobile Web Best Practices

¡  Device §  Slow CPU

§  10x slower then desktop §  JavaScript that runs for 100ms on desktop will take ~1s on mobile

§  Low memory §  Mobile RAM in MB §  Desktop RAM in GB

§  Small screen §  Navigation

§  Touch §  Gestures: Tap, Double Tap, Drag, Flick, Pinch, Spread, Press, Press & Tap, Press & Drag,

Rotate, … §  More direct interaction §  But different on each device

§  Focus §  Pointer

WHAT’S MOBILE – THE PROBLEM

Google Tech Talk - November 10, 2011 - Guy Podjarny

Page 7: Mobile Web Best Practices

¡  Software §  Rapidly changing, widely diverse browsers and operating systems

WHAT’S MOBILE – THE PROBLEM

No Browser Creator

1 Android Browser Google

2 Apollo Team Apollo

3 BlackBerry Browser Research in Motion

4 Blazer Palm

5 Bada Samsung

6 Chrome for Android Google

7 Classilla Cameron Kaiser

8 Deepfish Microsoft

9 Dolphin Browser MoboTap

10 Firefox for mobile Mozilla Foundation

11 ibisBrowser ibis

12 Internet Explorer Mobile Microsoft

13 Iris Browser Torch Mobile

14 JOCA InteracT!V

15 Kindle Basic Web Amazon.com

16 MIB Motorola

17 Minimo Mozilla Foundation

18 Myriad / Openwave Myriad Group

19 NetFront ACCESS

20 Nokia Series 40 Browser Nokia

No Browser Creator

21 Obigo Browser Obigo AB

22 Opera Mini Opera Software

23 Opera Mobile Opera Software

24 Palm HP

25 Pixo Sun Microsystems

26 PlayStation Portable Sony

27 Polaris Browser Infraware

28 Safari Apple

29 Symbian / S60 Nokia

30 Skweezer

31 Skyfire Skyfire Labs

32 Skyfire Mobile Browser Skyfire

33 Steel

34 Teashark

35 Tristit

36 UC Browser UC Mobile

37 uZard Web Logicplant

38 Vision Mobile Browser Novarra

39 WebOS Browser Palm

40 WinWAP Winwap Technologies

Page 8: Mobile Web Best Practices

¡  Software §  Changing rapidly §  Different navigation method

§  Proxied vs Direct §  Proxied – Opera Mini, Kindle Fire §  Direct – all others

§  HTTP vs SPDY §  SPDY – Kindle Fire, Chrome for Android §  HTTP – all others

§  Process models §  Multi Process – Chrome for Android §  Single Process - all others

§  Small Cache §  2-4MB

§  Connections §  Max Total §  Max per host §  Pipelined

§  Multiple requests over a single socket without waiting for a response §  Problems with proxies and head-of-line blocking

WHAT’S MOBILE – THE PROBLEM

Page 9: Mobile Web Best Practices

¡  Users §  Distracted

§  i.e. walking down the street

§  Shorter interactions §  i.e. waiting for a bus

§  Touch §  Fat fingers §  Gestures (are they all intuitive?)

§  Targeted tasks §  i.e. what are the local restaurants?

§  Expectations §  Web that “pretends” to be app is likely to disappoint §  85% - “customers expect to be able to shop on their phones and want the

experience to be as good or better than on a computer”

WHAT’S MOBILE – THE PROBLEM

http://www.nytimes.com/2011/04/18/technology/18mobile.html?_r=3

Page 10: Mobile Web Best Practices

¡ Mobile first §  Start simple and build up §  Test for presence (and function) of all APIs

§  If it exists it may not work correctly

§  Everything should fallback §  local storage → local variable → no storage §  push & pop state → hash bang fall → page refresh §  auto refresh → manual refresh button

§  Avoid native events → use synthetic events §  minimizes difference between browsers

¡  Forget pixel perfect §  Fluid grids §  Adaptive design

MOBILE FIRST

Page 11: Mobile Web Best Practices

¡  Avoid JS → Middle Tier § Why are we re-implementing the browser? § Maximize stability across devices, even unknown devices §  Ensure site works without JS §  You control execution environment

§  More reliable §  Easier to test

§  Limit device dependent code §  Easier to automate tests

§  Simplify browser interaction helps Selenium §  More code covered by Java unit tests

§  Easier to optimize §  Scale out / up §  Easier to diagnose issues §  Well understood §  Better tools

AVOID JAVASCRIPT

Page 12: Mobile Web Best Practices

¡  Breakpoints § Major breakpoint on server

§  Chose fixed high level breakpoints §  Mobile §  Tablet §  Desktop §  TV

§  Avoid serving desktop site on mobile width §  Too heavy §  Too slow to download §  Too slow to render §  Not optimized correctly

§  Adapt §  Content §  Layout §  Features §  Page weight

§  Caching §  Vary: User-Agent!

ADAPT – DEVICE DETECTION

Page 13: Mobile Web Best Practices

¡  Breakpoints §  How to select major breakpoint

§  Headers §  User-Agent §  Accept §  x-wap-profile / Profile

§  UAProfile with device capabilities and preferences (W3C Ubiquitous Web Applications Working Group – closed July 2010) §  OperaMini

§  Device Libraries §  WURFL

§  Open Source §  Slightly limited information §  Based on UAProfile

§  DeviceAtlas §  Started by Ericsson, Google, GSM Association, Microsoft, Nokia, Samsung, T-Mobile, Telecom, Telecom Italia, Visa and Vodafone

§  51Degrees §  Build into ASP.NET and MVC.NET

§  DetectRight §  Custom DB

§  Augment existing device library §  Log capabilities (i.e. cookie / analytics / JS)

§  Tools §  Spring Mobile

§  WURFL §  Apache Mobile Filter

§  WURFL §  51Degrees §  DetectRight

ADAPT – DEVICE DETECTION

Page 14: Mobile Web Best Practices

¡  Breakpoints §  Minor breakpoint on client

§  @media §  Screen size, resolution & pixel density

§  Fluid grids, dif ferent columns, adapt image size (see earlier slide) §  navigator.connection.type!

§  ETHERNET, WIFI, CELL_2G, CELL_3G, …!§  Send to server

§  Cookie §  JS

§  Landscape vs portrait §  Consider width of screen not width of viewport

§  Test for presence (and function) of all APIs §  If it exists it may not work correctly §  Everything should fallback

§  local storage → local variable → no storage §  push & pop state → hash bang fall → page refresh §  auto refresh → manual refresh button

§  Indicate it’s a mobile website

ADAPT – ADAPTIVE DESIGN

Page 15: Mobile Web Best Practices

¡  Progressive Enhancement §  Test for presence (and function) of all

APIs §  If it exists it may not work correctly §  Everything should fallback

§  local storage → local variable → no storage §  push & pop state → hash bang fall → page

refresh §  auto refresh → manual refresh button

§  HTML5 §  Not fully supported

§  mobilehtml5.org §  www.modernizr.com §  caniuse.com §  www.browserscope.org §  www.quirksmode.org

§  PolyFills §  Add weight §  Better to except not supported and degrade

ADAPT – PROGRESSIVE ENHANCEMENT

HTML

Simple CSS

Advanced CSS

Basic JavaScript

CSS Extensions

AJAX / HIJAX

HTML5

Page 16: Mobile Web Best Practices

¡  47% expect page load < 2s ¡  40% abandon sites with page load > 3s ¡  +1s → -7% conversion rate ¡  if $100,000 per day, +1s → $2.5 million lost sales every year ¡  Bing

§  +1s → -4% revenue ¡  Google

§  0.4s to 0.9s → -25% searches (-$2.5 billion revenue) ¡  Firefox

§  -2.2s download page → +15% downloads (+1.7 million extra / year) ¡  Shopzilla

§  -5s → +7-12% revenue, -50% hardware cost ¡  Wallmart & Amazon

§  every -100ms → +1% revenue

HIGH PERFORMANCE – WHY?

http://blog.kissmetrics.com/loading-time/ http://blog.mozilla.com/metrics/category/website-optimization/

http://radar.oreilly.com/2009/07/velocity-making-your-site-fast.html http://www.webperformancetoday.com/2012/02/28/4-awesome-slides-showing-how-page-speed-correlates-to-business-metrics-at-walmart-com/

Page 17: Mobile Web Best Practices

¡  Less Requests

HIGH PERFORMANCE – LESS REQUESTS

0.35

0.4

0.45

0.5

Total Requests Image Requests Total Xfer Size Image Xfer Size Domains

Highest Correlation to Load Time

0.3

0.35

0.4

0.45

0.5

Total Requests Image Requests Total Xfer Size Image Xfer Size Domains

Highest Correlation to Render Time

http://mobile.httparchive.org/

Page 18: Mobile Web Best Practices

¡  Less Requests §  Bundling

§  Maximum single request each for JS & CSS §  Single bundle more efficient use of bandwidth §  Blocks page while bundle downloaded §  Blocks page while bundle parsed §  Consider GMail approach to defer parsing (discussed later)

§  Image Sprites §  Single PNG image §  Data URI

§  Supported by: iPhone, Android, Opera §  Not supported by: IE, Nokia Series 40

§  spriteme.org, css-sprit.es §  CSS3 for effects & ornaments (not images)

§  Gradients, rounder corners, drop shadow, text shadow §  Avoid @font-face

§  Size is between 10KB up to 4-5MB for Asian character sets §  Flash of un-styled text (FOUT) due to download & install §  6% of mobile pages use @font-face

§  Aim at <= 1 requests §  We can do better than this….

HIGH PERFORMANCE – LESS REQUESTS

Page 19: Mobile Web Best Practices

¡  Less Requests §  Inline scripts, css and images on first load

§  Inline all resources, and send cookie §  Extract script, style and images into localStorage, update cookie §  Subsequent load check cookie

§  If initial value §  no JavaScript or localStorage - return page with resources as external links (to

utilize caching) §  If updated

§  Output script to load resources from localStorage (at top of page)

§  Bing §  1st request 54.9 KB §  2nd request 5.5 KB

HIGH PERFORMANCE – LESS REQUESTS

Page 20: Mobile Web Best Practices

¡  Less Requests §  Application Cache

§  <html manifest=“/manifest” >!§  Pros:

§  Extremely fast §  Second page load 0 HTTP requests (for static content) §  Authoritative control of locally cached resources

§  Cons: §  If file changes must update manifest file §  Makes page load asynchronous → second page load only §  Adaptive images → user cookie set on first load to determine image quality §  Double refresh issue → use updateready event on window.applicationCache and prompt user §  Atomic §  Impossible to version §  Buggy JS API §  Hard to invalidate

§  Tip: §  Loads in order so CSS → PNG → JS §  Use NETWORK: * to make manifest “open”

HIGH PERFORMANCE – LESS REQUESTS

Page 21: Mobile Web Best Practices

¡  Less Requests §  Application Cache

HIGH PERFORMANCE – LESS REQUESTS

Page 22: Mobile Web Best Practices

¡  Less Bytes

HIGH PERFORMANCE – LESS BYTES

http://mobile.httparchive.org/

0

100

200

300

400

500

Images Scripts Stylesheets HTML Other Total

Average Bytes per Page by Content Type (in KB)

0

2

4

6

8

10

12

14

16

JPEG PNG GIF HTML JS CSS

Average Individual Response Size (in KB)

Page 23: Mobile Web Best Practices

¡  Less Bytes §  Simple semantic HTML

§  No divitis, classitis, iditis §  However:

§  classes can improve css selector performance §  Ids can improve testability

§ Minify §  HTML, CSS, JS

§  Optimize Images §  Use PNG instead of JPEG and GIF

§  GIFs for small images (i.e. <10x10 pixels) §  JPEG for photographic-style images

§  Maximize lossless compression §  JPEG: jpegtran, jpegoptim §  PNG: OptiPNG, PNGOUT §  WebPageTest

HIGH PERFORMANCE – LESS BYTES

JPEG 47%

GIF 29%

Other 0%

PNG 24%

Mobile Image Requests

http://mobile.httparchive.org/

Page 24: Mobile Web Best Practices

¡  Less Bytes §  Adaptive Images

§  Don’t scale images on the client §  Wastes bandwidth §  Requires device CPU

§  Use @media, window.matchMedia(…), window.devicePixelRatio, etc §  -webkit-min-device-pixel-ratio (1 -> low res, 1.5 or 2 -> high res) §  For manifest file use capabilities cookie

§  Server Scaling §  Multiple fixed breakpoints

§  Allows for dif ferent designs (i.e. larger text on small images) §  Scale on demand in between major breakpoints §  Use CDN to “save” scaled images

HIGH PERFORMANCE – LESS BYTES

Page 25: Mobile Web Best Practices

¡  Less Bytes §  Less JS

§  Remove duplication §  Cost of parsing JavaScript about 1ms per 100KB §  Steve Souders – “every byte of JavaScript is 10x more expensive then

everything else” §  Mini frameworks only

§  Avoid → jQuery, Sencha, YUI, Dojo, etc §  It takes between 1.5s and 8s to parse and download jQuery over 3G §  jQueryMobile wraps jQueryUI with wraps jQuery Core

§  Use → XUI, Zepto, microJS, baseJS

HIGH PERFORMANCE – LESS BYTES

http://www.slideshare.net/mobile/Gomez_Inc/optimizing-web-and-mobile-site-performance-using-page-speed

Page 26: Mobile Web Best Practices

¡  Less Bytes §  Compression

§  More consistent data will compress better §  CSS key-value pairs in same order §  HTML attribute in same order §  Use lowercase (or consistent) casing wherever possible §  Use consistent quoting for HTML attributes (always single or always double) §  Minify JS, CSS and HTML

§  Avoid gzip for image or other binary files §  These files are already compressed

HIGH PERFORMANCE – LESS BYTES

Page 27: Mobile Web Best Practices

¡  Less Bytes §  Less Cookies

§  Remove / reduce cookies §  Cookie Free Domain §  Server-Side Cookies (ID only)

§  Store data on server side and only send unique ID §  HIJAX

§  Smaller page request §  Could also use Hashbang

§  Twitter accepts both https://twitter.com/#!/foo and https://twitter.com/foo §  Only works with JavaScript enabled §  Its an ugly hack and looks like one

§  History API §  history.pushState() & ‘popstate’ event

§  Don’t forget forward and backwards isn’t a page reload, it’s a page display §  Endless Scrolling Pattern

§  Only download subset §  Less bytes §  Much more complexity §  Only worth while for very large data sets §  Recycle elements that are scrolled off screen

HIGH PERFORMANCE – LESS BYTES

Page 28: Mobile Web Best Practices

¡  Bandwidth Efficiency §  Fit request into single packet (if possible)

§  Packet size ~1500 bytes

§  Parallelize downloads §  Avoid blocking HTML parsing

§  JS at the bottom

§  Avoid delaying resource download §  Avoid @import §  document.write(…)

§  Max requests per domain §  Domain sharding

§  Adds DNS lookup → >6 request per domain

HIGH PERFORMANCE – BANDWIDTH

Page 29: Mobile Web Best Practices

¡  Bandwidth Efficiency §  Defer JS requests

§  Progressive enhance so delay in JS is never a problem §  defer & async

§  download and parse not delayed §  tells browser there is no document.write(…)!§  async

§  executes after download §  execution order not guaranteed

§  defer §  executes after parsing

HIGH PERFORMANCE – BANDWIDTH

http://peter.sh/experiments/asynchronous-and-deferred-javascript-execution-explained/

Page 30: Mobile Web Best Practices

¡  Bandwidth Efficiency §  Defer JS requests

§  Downloading JS asynchronously §  XHR Eval §  XHR Injection §  Script in iframe §  Script DOM element §  document.write(…)

§  Load scripts after onload to guarantee loaded asynchronous download

HIGH PERFORMANCE – BANDWIDTH

Page 31: Mobile Web Best Practices

¡  Bandwidth Efficiency §  Eager loading

§  Download in background and store in localStorage §  Mobile Web: <a class="eager" … />!§  Chromium: <link rel=“prerender” … />!

§  Avoid bad requests (i.e. 404) §  7% mobile pages have (4xx) errors

§  Avoid redirects §  73% mobile pages have (3xx) redirects §  If you can’t avoid cache

§  Language redirect for Accept-Language (see details on caching later) §  Expires <= Date (to disable HTTP/1.0 caches that don’t support Vary) §  Cache-Control: max-age >= 1 year §  Vary: Accept-Language

HIGH PERFORMANCE – BANDWIDTH

http://mobile.httparchive.org/

Page 32: Mobile Web Best Practices

¡  Bandwidth Efficiency §  Dynamic Update

HIGH PERFORMANCE – BANDWIDTH

http://mobilehtml5.org/

Page 33: Mobile Web Best Practices

¡  Bandwidth Efficiency §  Dynamic Update

§  AJAX Polling §  Long Polling §  Server Side Events §  Web Sockets

§  JSON vs HTML fragments §  Trade-off

§  Less bytes on network §  Identical fixed network costs §  Much more complex JavaScript code that must work across multiple devices

§  HTML fragment approach → simple identical JavaScript universally across the site §  JSON → custom JavaScript for every item of updatable content

HIGH PERFORMANCE – BANDWIDTH

Page 34: Mobile Web Best Practices

¡  Less Latency §  Less server time

§  Execute backend calls in parallel §  Avoid synchronization

§  Threads §  DB (i.e. locking)

§  NoSQL §  Eventual Consistency

§  Use Futures to only block when absolutely required §  ETag → short circuit backend tasks §  Flush header while waiting for backend calls

§  Allows CSS to be downloaded on client in parallel with backend calls §  HTTP/1.1

§  Trailer header allows addition of Content-Length, Last-Modifier and ETag with Transfer-Encoding: chunked §  Servlet 3.0 does NOT support this!!

§  Under load scaling §  Avoid “too many” threads per request

§  Threads should not be blocked waiting (i.e. IO / locks) §  No conversation state

§  Must be managed across servers §  Session replication §  Sticky sessions

§  Some state exists but not held on web server §  Decouple and isolates problem

§  Separation of concerns §  Better encapsulation §  Improves opportunity for graceful degradation

§  Web Tier Persistence §  Transient User Data (i.e. shopping basket, web UI preferences)

HIGH PERFORMANCE – LESS LATENCY

Page 35: Mobile Web Best Practices

¡  Less Latency § Minimize DNS lookups

§  Mobile Web analytics request §  230ms DNS lookup §  76ms TCP + HTTP connection §  58ms server time

§  Flush Early §  Check flush isn’t buffered

§  Apache, Network (i.e. Netscaler), gzip §  Only works on HTTP/1.1 connections §  Uses Transfer-Encoding: chunked!

§  Caching §  Make pages cacheable

§  Avoid cookie or header driven content §  Except User-Agent, Accept-Language, Accept-Encoding

§  Use path variable instead of query parameters §  Most proxy caches do not cache resources with query parameters

§  i.e. only latest version of Squid fixes this §  Use fingerprinting to dynamically control caching

HIGH PERFORMANCE – LESS LATENCY

Page 36: Mobile Web Best Practices

¡  Less Latency §  Caching

§  Expires (HTTP/1.0) §  Date when resource is stale

§  Cache-Control (HTTP/1.1) §  Overrides Expires header §  max-age - how long resource is fresh (in seconds) §  public, private – where item is cacheable §  no-store – never cache §  no-cache – always revalidate §  must-revalidate – revalidate if stale §  s-maxage, proxy-revalidate – proxy servers settings

§  Vary (HTTP/1.1) §  Allows cached items to be varied by header §  If using disable HTTP/1.0 caches by Expires <= Date

§  Tips §  Maximize freshness → 1 year

§  Android clears cache based on length of freshness → >10 years §  Last-Modified

§  Last time resource modified §  If-Modified-Since – conditional GET §  If-Unmodified-Since – conditional PUT, DELETE

§  ETag §  Ensure consistency across multiple web servers §  Unique identifier for specific resource version §  If-None-Match – conditional GET §  If-Match – conditional PUT, DELETE §  Favor semantic “week validator” over byte level “strong validator” §  Week Validator enables short-circuiting the server HTML generation §  Trailer: ETag supports Transfer-Encoding: chunked (not supported by Servlet 3.0)

§  304 Not Modified §  Last-Modified §  ETag

HIGH PERFORMANCE – LESS LATENCY

Page 37: Mobile Web Best Practices

¡  Less Latency §  Caching

§  Not used enough

HIGH PERFORMANCE – LESS LATENCY

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

None t <= 0 0 < t <= 1 1 < t <= 30 30 < t <= 365 365 < t

Cache-Control: max-age (days)

http://mobile.httparchive.org/

Page 38: Mobile Web Best Practices

¡  Less Latency §  Caching

§  Browser §  Limited size §  Average mobile page size 550KB → < 7 pages to full reload

HIGH PERFORMANCE – LESS LATENCY

OS (Device) Component Total Last-Modified ETag Survives Power Cycle

Android 2.1 (Nexus One) 2MB 2MB Yes Yes Yes

Android 2.2 (Glaxay S) 4MB 4MB Yes Yes Yes

Android 2.3 (Nexus S) 4MB 4MB Yes Yes Yes

Safari, iOS 3.1.3 (1st-gen iPhone) 0b 0b No No No

Safari, iOS 3.2 (iPad) 26KB 282KB Yes Yes No

Safari, iOS 4.0 (iPhone 3GS) 51KB 1MB Yes Yes No

Safari, iOS 4.0 (iPhone 4) 102KB 2MB Yes Yes No

webOS 1.4.1 (Palm Pre Plus) 1MB ? No No Yes

http://www.yuiblog.com/blog/2010/06/28/mobile-browser-cache-limits/ http://www.blaze.io/mobile/understanding-mobile-cache-sizes/ http://mobile.httparchive.org/trends.php#bytesTotal&reqTotal

Page 39: Mobile Web Best Practices

¡  Less Latency §  Caching

§  Reverse Proxy (i.e. ISP) §  Vary: Accept-Encoding

§  many proxy caches do not detect Content-Encoding correctly §  Cookies

§  Generally public caches won’t cache pages with cookies §  If they do cookies will be shared

§  Edge / Gateway (i.e. CDN) §  Edge Side Include (ESI)

§  Dynamic content assembly in cache §  Add icons:

§  favicon.ico §  apple-touch-icon.png §  apple-touch-icon-precomposed.png

HIGH PERFORMANCE – LESS LATENCY

Page 40: Mobile Web Best Practices

¡  Less Latency §  Offline Storage

§  Single origin policy §  No freshness or cache eviction §  No versioning §  No request → very fast and works offline §  Manifest file

§  see earlier slides §  Web Storage

§  Local Storage (or Session Storage) §  key/value storage of strings §  Simple but no query language §  No locks or transactions §  No indexes or complex data structures §  Slow, synchronous → blocking §  Up to 5MB §  Very fast §  Synchronous §  Only accessible from JavaScript

§  Indexed DB §  Flat-file DB §  Hierarchical key/value persistence §  Basic indexing §  Fully asynchronous §  Supports indexed filtering and lookup §  DB level transactions & locking §  Up to 5MB, then prompt user for more §  Only accessible from JavaScript

§  Web SQL Database §  SQLite in browser §  select, insert, update, delete, joins, inner selects, etc §  Up to 5MB, then prompt user for more §  Standard dropped by W3C §  Slow §  Only accessible from JavaScript

§  File API

HIGH PERFORMANCE – LESS LATENCY

Page 41: Mobile Web Best Practices

¡  Less Latency §  Radio Resource Control

§  HEAD request every 2-3 seconds to keep cell connection live §  Drains battery and adds congestion so don’t poll forever

§  Page Beacons (Analytics) §  Make asynchronous

§  Inject 1x1.gif §  Separate domain

§  Avoid site cookies infecting request §  Avoid analytics cookies infecting site

§  304 response §  Maximize benefit

§  Page flow §  Device capabilities

§  @media / JS cookie §  User-Agent

§  Connection Type §  navigator.connection.type!

§  ETHERNET, WIFI, CELL_2G, CELL_3G, …!§  Carrier

§  Massive’s Operator Identification Platform

HIGH PERFORMANCE – LESS LATENCY

Page 42: Mobile Web Best Practices

¡  Faster Page Render §  Simple semantic DOM

§  For every element (i.e. in Chrome) §  Node in DOM Tree §  RenderObject in RenderTree

§  Efficient CSS selectors §  Avoid universal rules §  Don’t qualify id rules with tag names or classes §  Don’t qualify class rules with tag names §  Make rules as specific as possible §  Avoid descendant selectors (especially Tag or Universal) §  Apply rules at the highest level and allow inheritance §  Remove redundant qualifies

§  ID selector with class or tag §  Class selector with tag

HIGH PERFORMANCE – PAGE RENDER

Page 43: Mobile Web Best Practices

¡  Faster Page Render §  CSS in head

§  External or inline style blocks in body can cause reflow or FOUC §  Allows page to render progressively §  External stylesheets block rendering until downloaded and parsed

§  Use hardware accelerated CSS animations §  Gets GPUs directly involved with compositing §  Avoids slow video to system memory copies §  Less redraws of complete screen

§  only compositing layer redraw and recomposition §  Use CSS instead of JS animation §  Specify image dimensions

§  Avoids reflow after image downloaded §  Specify character set

§  Avoids browser auto detecting (and making / correcting mistakes) §  Timers

§  Lots of >1s timers §  Very few (i.e. one) <500ms timers

HIGH PERFORMANCE – PAGE RENDER

http://googlecode.blogspot.com/2009/07/gmail-for-mobile-html5-series-using.html

Page 44: Mobile Web Best Practices

¡  Faster Interaction §  Defer JS parsing

§  100ms per 1KB §  Make browser ignore code

§  Comments §  None script element

§  When needed use eval(…)!§  Avoid blocking single UI Thread

§  Long-running JS = Unresponsive UI §  Asynchronous

§  Use events §  Synthetic (via setTimeout in framework) §  Native

§  Script Yeilding §  setTimeout(function() {…}, 50)!

§  Different browsers have different timer resolutions / issues §  Short timeouts can significantly shorten battery life

§  setImmediate(function() {…})!§  Add to back of UI Thread Queue (i.e. no timer) §  Not supported yet (only in IE 10)

§  WebWorkers §  Asynchronous, doesn’t block UI Thread §  Data is serialized into and out of worker §  No access to DOM §  Separate execution environment

HIGH PERFORMANCE – INTERACTION

Page 45: Mobile Web Best Practices

¡  Faster Interaction §  Touch delay

§  Click event delayed between 300 and 500ms §  ontouchstart → ontouchend (with ontouchdrag within error margin)

§  Event Handlers §  Event handlers on outer container

§  Use event bubble §  Avoid lots of smaller event handlers

HIGH PERFORMANCE – INTERACTION

Page 46: Mobile Web Best Practices

¡  Faster Interaction §  Touch

§  Work directly with content §  Avoid button indirection §  Although gesture support currently limited

§  Limited touch events §  touchstart, touchmove, touchend

§  Fat fingers §  Fitt’s Law

§  difficulty ∝ distance, size

§  Large buttons §  Inactive margin

§  Right first time

HIGH PERFORMANCE – INTERACTION

Page 47: Mobile Web Best Practices

¡  Improve Perception §  Flush early §  Visual feedback to user

§  Button click changes color

§  Image Interlacing §  Visible image after only 1/64 downloaded §  Adds ~15%

§  Remove address bar §  While JavaScript is executing in background §  After onload

§  window.scrollTo(0, 1);!§  If page height < window height then increase page height first to allow scroll

HIGH PERFORMANCE – PERCEPTION

Page 48: Mobile Web Best Practices

¡  Debugging §  Chrome Internals

§  chrome://net-internals §  chrome://appcache-internals/ §  chrome://profiler/ §  For full l ist see chrome://chrome-urls/ §  Works on Chrome for Android

§  Firebug §  Desktop only

§  BlackBerry Browser §  Remote debugging §  Smartphone 7.0+ §  PlayBook

§  Web Inspector §  http://trac.webkit.org/wiki/WebInspector

§  Wienre §  http://phonegap.github.com/weinre/

§  DragonFly (for Opera Mobile) §  WebKit Remote Debugging

§  http://www.webkit.org/blog/1620/webkit-remote-debugging/ §  Adobe Shadow

§  Synchronize desktop browser with multiple mobile browsers §  http://labs.adobe.com/technologies/shadow/

TOOLS – DEBUGGING

Page 49: Mobile Web Best Practices

¡  Performance / Analysis §  Chrome Inspector

§  JS & CSS Profiler §  PageSpeed

§  Chrome Inspector §  Firebug

§  YSlow §  Chrome Inspector §  Firebug

§  Mobile Perf Bookmarklets §  YSlow and other tools §  http://stevesouders.com/mobileperf/mobileperfbkm.php

§  PCAP Web Performance Analyzer §  http://pcapperf.appspot.com/ §  Analyse packets → Waterfall & WebPageTest

§  Shark for Root §  Android packet sniffer §  Works with 3G §  Analyse results (PCAP file) with pcapperf

§  Speed Tracer §  Chrome Inspector Plugin §  Analyses all page interaction (CSS & JS)

§  3PMobile §  On device browser §  Network Waterfall §  Browser Event (i.e. image loading time, CSS processing time)

§  dynaTrace §  JS performance analyser §  http://ajax.dynatrace.com/

§  Loadtimer §  Tests time for full load of multiple URLs §  http://www.stevesouders.com/blog/2011/12/01/loadtimer-a-mobile-test-harness/

§  WebPageTest §  Waterfall of network traffic §  Timing of page load event adds noise

TOOLS – PERFORMANCE

Page 50: Mobile Web Best Practices

¡  Emulators §  www.mobilexweb.com/emulators

¡  Remote Labs §  DeviceAnywhere (bought by WebPageTest)

§  deviceanywhere.com §  PerfectoMobile

§  perfectomobile.com §  WebPageTest §  Samsung

§  Free hosting of Samsung devices §  http://developer.samsung.com/remoteTestLab.do

§  Nokia §  Free hosting of Nokia devices

¡  Experimental §  Cuzillion

§  Build up different types of page §  internal / external CSS §  internal / external JS §  images

§  Good for testing and learning about new devices / browsers §  http://stevesouders.com/cuzillion/

TOOLS – EMULATORS