dotMobimobiThinkingmobiForgemobiReadyDeviceAtlasgoMobi
background image

Measuring the Effects of Operator Transcoders

ronan's picture
Posted by ronan - 23 Jun 2010
Twitter share icon Facebook share icon Google Plus share icon

The issue of mobile-operator transcoders on the mobile web has been hugely controversial. For many iPhone and Android phone owners the negative effects of the transcoders are no longer visible and the issue has dropped off the radar, but they remain a serious problem for those using lower-end devices.

The potential benefits of transcoders are clear: they can make some desktop-oriented web content work acceptably well on mobile devices. But the cost of this can be very large: many mobile operators are still utilizing transcoders that have disastrous effects on made-for-mobile content, without offering content owners any way to bypass them. This reduces usability and hampers the growth of the mobile web by disincentivizing the content creators.

We have now heard so much about this issue that we are considering taking further action.

Our approach will be a project to document the impact of operator transcoders around the world. We will build a set of test sites and pages, partially based on the W3C CTG tests, that will enable people to test what the transcoding engines are doing in different test scenarios. This will documented in a series of steps that should be performed. This is a very complex area since it depends on many variables:

  • the browser in question
  • the device in question
  • the page content and URI of the page under test
  • the APN in use by the device

To produce meaningful results all of the factors mentioned above will need to be documented for each result, since the tests are meaningless otherwise. It seems likely that, at most, the results of any test will say something like: For device X, provisioned with configuration Y, and APN Z, certain tests failed.

Human testing will be necessary because many of the searched-for effects are not programmatically discernible.

By necessity, this will need to be a community project, since we will need volunteers around the world to test their local operators. If we receive sufficient interest from the community we will proceed with this project in an open way. The intention is to use mobiForge as a central point where the effects of operator transcoders are documented. Our hope is that by raising awareness of the effects of transcoding and detailing who exactly is doing it badly, we can effect some changes.

Please let us know via comments or email if you think that this is a project worth pursuing, and if you would like to contribute via a forum to flesh out ideas.


Posted by ronan - 23 Jun 2010

ronan's picture

« m o b i l i s t , CTO of dotMobi » @xbs

Posted by screaminggeek 4 years ago

I applaud dotMobi for taking up the cause of transcoder awareness. Transcoders greatly impact unsuspecting mobile subscribers but enjoy relative unawareness among mobile web developers, especially in the iPhone- and Android-centric smartphone web development community.

The CTG tests measure compliance with one behavioral standard for transcoders, but they do not provide an adequate measure of how transcoders ultimately affect the user experience of the Web on mobile devices. I suggest adding tests that measure the usability of transcoded Web 2.0 apps. In addition to decomposing a Web page into an unnavigable number of fragments, markup transformation breaks AJAX, JavaScript and even basic Web functionality like form submission. dotMobi's evaluation criteria should test the preservation of Web usability as well as syntax.

dotMobi should consider testing using the criteria from the Manifesto for Responsible Reformatting from the WURFL community. Its common-sense criteria differs from the CTG but is adopted by some transcoders and advocated by many mobile developers.

Please also consider testing public transcoders. Google Web Transcoder and others can degrade the Mobile Web experience when featurephone users click on transcoded search results. I'm convinced that clickthrough rates decline when users learn that a SERP provides links to a transcoded Web experience (this is my intuition, I don't know of research in this area).

Thanks, hope this research gets off the ground with community involvement.

Gail Rahn Frederick
Learn the Mobile Web

Posted by belendo 4 years ago

I'll be happy to help with this, but I believe it's important to take into account Gail's comments. A research approach that incorporates both the CTG and the WURFL community manifesto might help reach a unified position towards transcoding, supported by the whole mobile community.

Posted by ronan 4 years ago

Gail,

Thanks for mentioning the Manifesto -- I should have mentioned it in the post.

The effects of the transcoders on Web 2.0 apps is a good point also -- now that the web has evolved beyond a simple declarative medium to one that requires the client to run code the landscape has become even more complex.

Hopefully some other folks will weigh in with support and we'll have enough momentum to get something done.

Ronan Cremin, dotMobi