Big Android Browser Test.
The purpose of this test is to compare various web browsers. It is not intended to be purely scientific but shall we say, significantly more scientific then unscientific ;).
To ensure completeness, I will have my test procedures fully documented here but am to lazy to YouTube it all. I desire very much a uniform Oh-My-Gosh-Test-Battery to attack each application with, and record the results.
- ASUS TF201-B1 Transformer Prime.
- ASUS TF101-B1 Transformer and Dock.
A dock for the Prime will not be arriving until sometime in Feburary (at best), so testing is likely to be in progress or finished by then.
My original Transformer has some touch screen issues, and I may need the dock. It is also how I will generally be testing “Mouse” related issues. The device also runs a custom ROM and I’m not going to go through the trouble of wiping it and reverting to stock.
Browser Applications Tested.
- TF101 = Honey Comb, Prime 2.1.1 ROM by roach2010.
- TF201 = Ice Cream Sandwich, stock ROM with latest firmware and rooted.
- Mobile version with “Turbo” caching-proxy service disabled.
- Mini version.
Web Sites Tested.
U.S. desktop versions will only be tested when possible.
Browser Under Test
Site Under Test
Device Under Test
User Agent string
To be done before starting tests, not before each test.
- Ensure all browser applications are installed.
- Restart the DUT.
- For the very first restart only: check for the “Running” applications in “Settings” to see if any browsers have installed services that start on boot.
- Place the device on charge.
Note that for the TF201, this also means tests are done with power saving mode set to “Balanced”. I will have my screen on auto brightness (usually pretty low), since this isn’t perfect rocket science.
Unless otherwise noted under “Specific.” this is how web sites were tested:
- Erase SB cache in browser settings and exit SB.
- Launch BUT under stopwatch.
- If BUT was previously used.
- Erase BUT cache in browser settings.
- Confirm stable configuration.
- Use “Exit” button / method if provided.
- If supported then set UA to Desktop.
- If stock bookmarks visible then test relationship to stock bookmark.
- Add bookmark.
- Test if visible in stock browser.
- Remove bookmark.
- Confirm no longer visible in stock browser.
- Load SUT in BUT.
- Switch to this document in Google Docs application and click link under “Web Sites Tested.” Scroll the list so that the BUT is visible.
- Or type that same URL into the address bar of the BUT.
- Start stopwatch and (if a. then click icon for BUT; else if b. then “Go”).
- Scroll page top to bottom.
- Scroll page bottom to top.
- Scroll page left to right if applicable.
- Scroll page right to left if applicable.
- Perform site specific test procedures OR explore a few links and document what was clicked.
- On browser test completion, “Exit” the BUT if a method is provided.
- If a “Exit” method is provided: check “Running” applications under “Settings” to ensure no associated service was left running.
For each site tested, steps 4 to 11 must be repeated.
For each browser tested, steps 1 amd above must be repeated.
- Search for C++ Programming.
- Open first link in new tab (where applicable).
- Search for images of Spiders.
- Load gmail.
- View message.
- Open draft (test) message for composer.
- Browse contacts.
- Search for DUT in shopping.
- Translate “Potato”
- to German and back.
- play German audio sample.
- play English audio sample.
- Load calendar.
- Create test event.
- Delete test event.
- Browse albums.
- Upload photo to dropbox album.
- Delete said photo.
- Open and edit a test document briefly.
- Open and edit a test spreadsheet briefly.
- Open and edit a test image briefly.
- Browse Groups.
- Browse Reader.
- Open a (test) draft post and edit.
- Post to stream about the current BUT.
- Browse around my local area.
Build Number: IML74K.US_epad-220.127.116.11-20120117 / PAD EC version: TF201-PAD-0214
 I use the in draft setting
 Now known as Freecode :-/.
 I’ll test this and probably test a private one as well.
 For a truly scientific study this would have to also be stripped down to a freshly setup STOCK+unrooted system at the firmware level under test.
 This hopefully avoids the impact of charging/battery level on the test results.
 I usually use Opera Mobile.
 Like making sure Opera Turbo is turned OFF in Opera Mobile.
 So that time to find the bloody thing doesn’t contaminate the timings.
 I really should learn how to use “Presentations” someday >_>.