Re: [wmlprogramming] Interesting article
- Dejan Kozina wrote:
> Just found this one, written by an Opera developer:I'll get the ball rolling by pointing out what appears to meas fallacies
> Seems better than average to me.
in the article:
- I am not familiar with Orkut, let alone with their mobile web version.
I have no doubt believing that it was so bad that users actually
preferred a transcoded version. But come on, that's just one
application. Everybody here knows that a typical mobile application
(even when relatively poorly coded) will always win hands-down when
compared to any transcoded web version in terms of usability. There is
no discussing it. You may come up with a scenario where the web site has
a function which has been left out in mobile, but trying to demonstrate
anything else is like trying to demonstrate that a SUV is preferable
when looking for a parking spot in the center of Rome.
- "77% of the traffic is reformatted full-web": this only proves that
users know that full-web is better when reformatted through OperaMini.
Users are using their regular browsers for sites which they know to work
OK on mobile. Also, a full-web page is a lot of reformatting (i.e.
higher number of reformatted pages per site) because of people clicking
around to find what they need. Mobile-optimised sites will typically
make people find what they are looking for when they are mobile sooner
(i.e. less clicks,i.e. less transcoding). In short, taking this 77% as
an indication that mobile surfing is less popular is wrong.
- It does not mention WURFL: I think this is a deadly sin for someone
who writes about the different ways to serve mobile content in 2009.
This is particularly true if you consider that Opera uses WURFL to
decide which versions of OperaMini to provide to downloaders. This
ommission seems intentional to me, particularly if you consider the
other non-omitted obsions in the list:
- "Mobile Device Detection and Style Sheets without User Agent Detection
or Server-Side Scripting
This is taken by the author for granted. Unfortunately for him, he
refers to a poorly-informed blog by someone who has not understood the
AdMob Metrics report (or, more simply, was desperately looking for stats
to bend evidence to his point in the face of reality).
To make a long story short, it is not true that 50% of deices around can
use CSS without UA detection/Server Side scripting. The AdMob metrics
has a graph about device OSes where a deviceOS is a development platform
for developers (i.e. new apps can be installed on the device). This
means that most feature phones (with simple, non-accessible, proprietary
OSes) are not accounted for. Yet, those devices are 85%+ of the whole
market! Someone hasn't done his homework here.
Of course, let's not forget that Opera is the producer of OperaMini,
i.e. a transcoding proxy. It is only obvious that their message is
"people want the full web", because they are making money on transcoding
as much as they can (they partnered with ByteMobile last year). By now,
we are not so easily fooled by transcoders, are we?