A few words about:

1
A few words about

• Browser incompatibility

2
Browser Incompatibility

• Imagine if every different TV set required a
  different signal to receive your favourite TV
  program - it sounds ridiculous, but that's close
  to the situation on the Web because browser
  makers have failed to implement a common set of
  standards.
• George Olsen (chair of the web standards group)

3

• What causes browser incompatibility?

4
Example (Windows) Browser Compatibility
Chart source http//www.webmonkey.com
5
Degree of support

• Although in some cases several browsers support a
  particular feature (e.g. JavaScript) - often the
  extent or nature of that support is different
  between different browsers, different versions of
  the same browser and/or the same version of the
  same browser on a different platform

6
Importance of testing

• As many of these cross-browser issues are very
  complex it is highly likely that you will not
  even realise that your site will look odd in
  certain browsers.
• This serves to highlight the importance of
  testing with several browsers/platforms, either
  artificially by the author - or better still by
  typical prospective users using their preferred
  browsers.

7
Automated Tools
8
What can be done about this?

• In the long term we can hope that manufacturers
  standardise core features of their browsers.
• In the short term however we can
• produce sites that attempt to use only the most
  commonly used HTML elements (and plugins,
  interactive facilities etc...) in the hope that
  the site will be visible on all systems.
• produce several versions of a site, optimised for
  different browsers.
• conduct an analysis of our sites user base in the
  hope we can optimise a site for their preferred
  browsers.

9
Detecting a browser

• If you do create multiple versions of your site
  it is desirable to be able to detect the users
  browser and provide them with the appropriate
  version.
• Each browser will send its User-Agent as part of
  each HTTP request it makes,e.g.
• User Agent Mozilla/4.0 (compatible MSIE 4.01
  Windows 98)
• Sadly there is little standardisation of how
  these User Agent strings should be constructed
  making the job of interpreting the browser type
  difficult.
• A popular (although flawed) technique for
  identifying the browser type is to run a client
  side JavaScript to work out the browser type -
  works fine unless the particular browser does not
  support JavaScript (or has it turned off).

10

• What could you do to minimise browser
  incompatibility?

11
Simple advice

• A straightforward technique for enhancing your
  cross-browser compatibility is to avoid the most
  recent features provided by a browser - these are
  often subject to the most change and least likely
  to be supported by competing browsers.
• Sadly of course on occasions your project may
  depend on these features
• If you must use recent features - tell the user
  and possibly provide download links to relevant
  resources.

12
Some of the most common browsers
13
Browser Statistics

• If designing a site for a hypothetical typical
  web user - it is useful to know what browsers
  will account for the majority of the average web
  sites traffic - many sources publish such
  information on the web itself.

14
Source 1 http//www.w3schools.com/browsers/browse
rs_stats.asp Source 2 http//www.webreference.
com/stats/browser.html
15
Some problems with browsers statistics

• most browser stats are based on visits to only
  certain sites - and the users of those sites may
  not be representative of the average user.
• It is not trivial to identify which browser a
  user is using, as many browsers will report
  themselves as either Netscape or IE type browsers
  and be misidentified.
• Statistics need to be current to be of much use
  to you - things can change very quickly.
• Sometimes more detail is needed than is provided
  (e.g. browser version).

16
Browserwatch Statistics
statistics represent visitors to
http//www.browserwatch.com which represents
users with an interest in browsers, a
significant portion of which are employed
professionally in the web authoring business.
17

• These tables summarise browsers used by 9819
  hosts in 67 countries making 98043 accesses on
  26/Feb/2002 to the Engineering Workstations WWW
  server at the University of Illinois at
  Urbana-Champaign.

2002
18
Chicago browser detail
19
A bit of history browser war

• It is relatively recently that Internet Explorer
  has been the most popular browser - IE was not
  released until several years after the web became
  popular.
• Previous popular browsers included Netscapes
  Navigator - and its predecessor NCSA Mosaic

20

• The tables below summarise which browsers were
  used by the 29910 hosts making 85197 accesses in
  the last 9.3 days (from Wed 10 Jan 1996) to the
  Random Yahoo Link, a CGI script that redirects a
  browser to a random link chosen from the tens of
  thousands in the Yahoo database.

1996
21
Wider issues

• Some problems that are related to browser
  incompatibilities are not, often we merely wish
  our site to appear as we intended on the users
  machine - other matters to consider include
• the screen space the user has available
  (resolution)
• the plugins the user has installed
• the colour depth of the users monitor
• sound reproduction capabilities of the users
  machine
• network bandwidth available to the user
• the platform/operating system that the user has
• the preference settings the user may have made
  for a browser

22
Monitor Resolution Stats

• Sourcehttp//www.upsdell.com/BrowserNews/stat_tre
  nds.htm

Monitor Resolution (pixels) Observed
------------------------------------------------
------ 640x480 (VGA) lt0.2 800x600
(XGA) 10 1024x768 (SVGA) 82 and
better other significant resolutions are 544x372
585x386 at 2 which are TV based browser
resolutions.
23
Plugin UseSource www.statmarket.com
(commercial provider, year 2000 data.)
As you can see on this basis use of any plugin
will exclude at least 1 user in 4, for many
plugins the picture looks even worse (of course
your users may be different!)
24
Bandwidth

• Bandwidth is clearly significant- often as a
  designer you will have high speed access to your
  files, users probably will not. Here is
  statmarkets estimate of the users that will
  remain at a page after seconds waiting for it to
  load

25
Some examples
26
Further reading / activities

• http//www.netmechanic.com/products/Browser-Tutori
  al.shtml
• http//www.upsdell.com/BrowserNews/index.htm
• http//www.quirksmode.org/dom/compatibility.html
• Search for screen resolution statistics and
  compare the results

27
Discussion

• Why would anyone develop an incompatible
  browser?
• Are there any advantages for web designers?
• What can we do about browser compatibility?
• Firefox vs IE (community vs Corp)
• Anything else?