Some stats on the Web

Miscellaneous Topics
#1

3DzzD have few visitors in comparaison of other websites but using analitycs on a cupple of webpage it grab some usefull informations, here are some statistics for last month that can be usefull for Web development.

would be really cool and usefull if someone can post similar one but with a higer traffic, as this one may not be very smart ?

so a good target audience could be Java with IE/FF on Windows for a screen higer or equals to 1024*768 using 32bit

OS:

  1. Windows 85,88 %
  2. Macintosh 8,54 %
  3. Linux 5,48 %
  4. FreeBSD 0,05 %
  5. SunOS 0,05 %

Java (not sure google detect 1.1 ?):

  1. Yes 91,75 %
  2. No 8,25 %

Flash:

  1. 9.0 r124 72,30 %
  2. 9.0 r115 10,92 %
  3. 9.0 r47 3,93 %
    4. (not set) 3,78 % probably no flash ?
  4. 9.0 r45 3,64 %
  5. 9.0 r28 1,02 %
  6. 10.0 b218 0,87 %
  7. 8.0 r22 0,53 %
  8. 10.0 r12 0,49 %
  9. 7.0 r19 0,29 %
  10. 9.0 r16 0,29 %
  11. 9.0 r48 0,29 %
  12. 10.0 r2 0,19 %
  13. 10.0.0 d525 0,19 %
  14. 8.0 r24 0,19 %
  15. 10.0 r1 0,15 %
  16. 10.0 r0 0,10 %
  17. 4.0 (GPLFLash) 0,10 %
  18. 9.0 r19 0,10 %
  19. 10.0 d26 0,05 %
  20. 10.0 d525 0,05 %
  21. 10.0 d536 0,05 %
  22. 10.0.2 d26 0,05 %
  23. 6.0 r79 0,05 %
  24. 6.0 r88 0,05 %
  25. 7.0 r25 0,05 %
  26. 8.0 (Gnash) 0,05 %
  27. 9.0 (Gnash) 0,05 %
  28. 9.0 r100 0,05 %
  29. 9.0 r125 0,05 %
  30. 9.0 r31 0,05 %
  31. 9.0 r64 0,05 %

Browsers/OS :

  1. Firefox / Windows 49,01 %
  2. Internet Explorer / Windows 29,79 %
  3. Firefox / Macintosh 4,42 %
  4. Safari / Macintosh 3,78 %
  5. Opera / Windows 3,74 %
  6. Firefox / Linux 3,15 %
  7. Chrome / Windows 2,47 %
  8. Mozilla / Linux 1,80 %
  9. Safari / Windows 0,63 %
  10. Konqueror / Linux 0,34 %
  11. Opera / Linux 0,19 %
  12. Opera / Macintosh 0,19 %
  13. Mozilla / Windows 0,15 %
  14. Mozilla / Macintosh 0,10 %
  15. Camino / Macintosh 0,05 %
  16. Firefox / FreeBSD 0,05 %
  17. Firefox / SunOS 0,05 %
  18. Netscape / Windows 0,05 %
  19. SeaMonkey / Windows 0,05 %

Screen bit depth:

  1. 32-bit 83,02 %
  2. 24-bit 11,84 %
  3. 16-bit 5,14 %

Screen resolution:

  1. 1024x768 26,01 %
  2. 1280x1024 22,61 %
  3. 1680x1050 12,32 %
  4. 1280x800 12,18 %
  5. 1440x900 8,78 %
  6. 1920x1200 5,19 %
  7. 1400x1050 2,52 %
  8. 1152x864 2,13 %
  9. 1600x1200 2,13 %
  10. 800x600 1,26 %
  11. 1280x768 0,97 %
  12. 1280x960 0,73 %
  13. 1024x600 0,53 %
  14. 2560x1600 0,39 %
  15. 1360x768 0,29 %
  16. 1600x1024 0,24 %
  17. 1536x960 0,19 %
  18. 1280x720 0,15 %
  19. 1344x840 0,15 %
  20. 1152x720 0,10 %
  21. 1366x768 0,10 %
  22. 2560x1024 0,10 %
  23. 832x624 0,10 %
  24. 969x768 0,10 %
  25. 1050x1680 0,05 %
  26. 1152x870 0,05 %
  27. 1200x1600 0,05 %
  28. 1280x638 0,05 %
  29. 1280x854 0,05 %
  30. 1360x1024 0,05 %
  31. 1440x852 0,05 %
  32. 1600x900 0,05 %
  33. 1680x945 0,05 %
  34. 1920x1080 0,05 %
  35. 2560x800 0,05 %
  36. 3840x1024 0,05 %
  37. 747x597 0,05 %
  38. 768x1024 0,05 %
  39. 960x600 0,05 %

Countries:

  1. United States
  2. France
  3. Germany
  4. United Kingdom
  5. South Korea
  6. China
  7. Spain
  8. Lithuania
  9. Russia
  10. Canada
  11. Italy
  12. Netherlands
  13. Switzerland
  14. India
  15. Belgium
  16. Brazil
  17. Australia
  18. Finland
  19. Poland
  20. Sweden
  21. Thailand
  22. Hungary
  23. Hong Kong
  24. Philippines
  25. Romania
  26. Ireland
  27. Turkey
  28. Tunisia
  29. Czech Republic
  30. Greece
  31. New Zealand
  32. Denmark
  33. Mongolia
  34. Malaysia
  35. Israel
  36. Portugal
  37. Colombia
  38. Ukraine
  39. Indonesia
  40. Slovakia
  41. Mexico
  42. Japan
  43. Taiwan
  44. Senegal
  45. Austria
  46. Singapore
  47. Norway
  48. Vietnam
  49. Egypt
  50. Bulgaria
  51. Pakistan
  52. Slovenia
  53. Croatia
  54. Serbia
  55. Uruguay
  56. Argentina
  57. South Africa
  58. Peru
  59. Algeria
  60. Iran
  61. Chile
  62. Morocco
  63. Saudi Arabia
  64. Bangladesh
  65. United Arab Emirates
  66. Reunion
  67. Puerto Rico
  68. Malta
  69. Latvia
  70. Monaco
  71. Brunei
  72. Cyprus
  73. Macedonia
  74. Iraq
  75. Kenya
  76. (not set)
  77. Iceland
  78. Bolivia
  79. Uzbekistan
  80. Venezuela
  81. Mauritius
  82. Moldova
  83. San Marino
  84. Luxembourg
  85. Tanzania
  86. Georgia
  87. Palestinian Territory
  88. Albania
  89. French Polynesia
  90. Belarus
  91. Sri Lanka
#2

You seem to target mainly Windows… If you want, I can post my figures, I have about 1400 distinct visitors per month.

#3

If this is something you’ve written yourself, consider switching to Google Analytics (which is what I use). It’s free and very easy to integrate and generates very robust stats. Pretty graphs too as a bonus. :slight_smile:

#4

Google Analitics calls JavaScript: Java

so… >:(

#5

those are already analitycs results, but I found that it is also good to merge with serverside (as analitycs is client side)

[quote]Google Analitics calls JavaScript: Java
[/quote]
is it the test used for Java enable, this will mean not working on some browser as opera right ??

[quote]You seem to target mainly Windows… If you want, I can post my figures, I have about 1400 distinct visitors per month.
[/quote]
this will certainly be interresting to see what is common (I do not target window user or anyothet, no special target at all, there are just less linux user than window I guess), but I would also like to see a higger traffic as something like more than 100000 users per month (dzzd.net do not have a lot of visitors but there is already some for a so specific web site, I got my maximum on july arround 12000 and 41500 page viewed (thanks to nehe…) and this year dzzd.net got around 3000 visitors and 20000 pages view per month).

#6

There is simply no reliable way to determine whether Java is available in the browser,

other than embedding an applet that sends a message to the server which is logged (at which point Firefox screws up)

I came to this conclusion after months of using various approaches. window.navigator.javaEnabled is a joke: it reports whether you-have-not-forcibly-disabled-java (in some obscure properties window in MSIE). So it reports true when-you-did-not-disable-it, regardless of whether Java is installed.

On quite a few freshly bought laptops, I have installed Java, but it wouldn’t work, as Java-was-forcibly-disabled by default, right out of the factory.

MSIE > extra > internet properties > tab[programs] > button[manage plugins] > find Java Plugin, and toggle the radiobutton to ‘enabled’. Right… like anybody is going to find it, I pulled a few hairs!

So now that all fancy advanced javascripts that search for the JRE plugins report that Java is installed, it still won’t work. Worse, quite some browsers report they don’t have Java (either with trivial scripts or advanced stuff), but if you embed the applet, it gets launched (hurray!).

If you use <applet …> the message you get when you do not have Java, is horrible (and in MSIE even not pointing you to a download, but to the above mentioned obscure properties window). You really MUST use and , which breaks in Firefox 3.0 (why!) and Safari - at least you know you have Java installed with Mac OS X.

So… anybody that collects statistics of installed Java versions without launching an applet and reporting back to the host server, has a massive error-margin.

Just a few stats of mine:
88% of the visitors’ browsers of a non-tech-related site, report they have Java.
54% can launch the applet in 3 tries or less. (byebye customer)

Yes, even if you have Java (even 1.6.0_xx), the plugin tends to get into deadlock before your code is executed, or crashes on creating the bright orange preloader with some GraphicsConfiguration.createCompatibleImage problem (yes…) and then you got the Firefox plugin problems that prevent you from communicating with your hostserver.

I work for a company where I convinced my boss to use a Java applet (over javascript + DOM manipulation). That application is the main way to make customers customize their products, and thus keep the company going. After 6 months of struggling (!) we had to resort to a ‘standalone version’ of the applet (13MB, JRE embedded).

The Java Plugin user experience is CRAP, I could have learnt Flash in the meantime and finish a product that worked everywhere.

#7

arf … :’( java web user experience is worse than what I was thinking…
really sun MUST do something to improve the user experience for Java on the Web, if flash improve its language (and be sure flash is not stupid and that accelerated 3d will be on soon…) a day Java will totally disapear (for web client side) !!!

NB: I just think of a way that could work to detect java, this way ?

with hasJava.php logging that’s the user have Java and return the readfile of the file hasJava.php.class

#8

Hm… nice idea. I bet however that you must provide a MyClass.class file, which is a resource, not a file (with apache mod_rewrite for example).

Maybe it would work… (wanna try?)

About Sun fixing the Java Plugin… they have always focused on serverside / standalone apps. Only recently with 6u10 we see some improvement. So lets be optimistic and say 6u11 would be perfect… it would still take at least 5 years before 6u11 would have that penetration of >80%.

Java 1.1 and 1.4.2 are still very widespread, and will be for the coming years. They will only vanish when those computers vanish. After 1.5 the nagging ‘update available’ will bring most users the latest and greatest.

#9

hum wait you are right… a class ending with .php will fail but he! will already have the information we want to :), we just have to do the following server side on the apache logs:

grep hasJava.php | wc //to know number of user requesting the page with the applet
and
grep hasJava.class | wc //to know number of user with java enabled

no?

#10

You’re right, but…

don’t forget <applet…> either triggers a very nasty dialog in MSIE, and an ‘information page’ that does not clearly tell you that you have to download Java, but tells you you have to change something in that obscure Plugin-management-window. So if you want to do it like that, you will basically confuse a lot of people.

And if java is installed, launching that applet for the first time will freeze your browser for a few seconds.

That is only just acceptable on a Java gaming site, I guess :slight_smile: On a commercial site it won’t be an option, although it would have given me some more information.

Anyway, you can make: /some_path_not_a_directory/MyApplet.class
which mod_rewrite rewrites to /my_binaries/MyAppletProvider.php
which writes all kinds of stats in the database, and then transfers /my_binaries/MyApplet.class

Now that you have an applet running, you can just as well connect to the host and tell the JRE version number, and put that in your database too.

#11

If you’re interested in the stacktrace of the Orange Preloader:


Exception in thread "Thread-27" java.lang.NullPointerException
	at sun.plugin.util.AnimationPanel.createTranslucentImage(Unknown Source)
	at sun.plugin.util.AnimationPanel.createGradientShapeImage(Unknown Source)
	at sun.plugin.util.AnimationPanel.initBackground(Unknown Source)
	at sun.plugin.util.AnimationPanel.preloadResources(Unknown Source)
	at sun.plugin.util.AnimationPanel.doPaint(Unknown Source)
	at sun.plugin.util.AnimationPanel.run(Unknown Source)
	at java.lang.Thread.run(Unknown Source)

Applets hangs.

#12

yup, I got that too sometime randomly, any idea ? workaround ?

#13

Huh? Works fine for me.

Oh, and if the .jar file pointed at by the tag is requested, the user probably has some kind of java installed.

#14

In very specific situations the Firefox plugin throws AccessControlExceptions when connecting to your own host.

This is mainly the case when the first connect attempt is not on the EDT. So the workaround is to ‘download’ a tiny file (few bytes) from your host in the Applet.init() method. After that you can connect to your own host from every other thread too.

#15

Cute. Thanks for the heads up, this is valuable information. =)

#16

I nearly pulled my hear out before I came up with this solution.

It was either a workaround or losing 20% of our costumers. >:(

#17

If I’m not mistaken the AnimationPanel is avoided entirely if you specify a background image.
http://java.sun.com/j2se/1.5.0/docs/guide/plugin/developer_guide/special_attributes.html#image

#18

MMM it suprised me that more people were using firefox than internet explorer!

By the way, what is the most popular version of jre used by people?

#19

yes I also think this is a strange result too, but I got nearly the same results serverside with a completly different analyser. on some other websites I got rather 20% FF VS 40 % IE6 and 25% IE7

#20

I get 56.3 % of users with Firefox, 21.1 % users that launch TUER via JavaWebstart directly, 11.8 % with Internet Explorer, 3.5 % with Mozilla, 3.2 % with Safari, 2.7 % with Opera, 0.5 % with Konqueror, 0.1 % with Netscape and 0.2 % with an unknown browser. TUER might get more than 3000 different visitors this month which would be the best result since the beginning of the project in october 2006.