Tag: Mobile Web

Peculiar device recognition by Disney

Earlier today I followed a link on mocoNews about Disney announcing record visits on their mobile site. I remembered Disney has a nice disney.mobi address, so when I read that the article talked about visitors to “disney.com” I was a bit surprised and wondered if it was a mistake. With my Sony Ericsson V640i I went straight to the .com and to my great surprised I saw a nice mobile site. My curiosity immediately grew from 0 to 100!

I opened a nice xterm and started playing with curl and headers. Here are a few discoveries:

  • disney.go.com is the real address, no matter what you use, you are always using the .go.com domain

  • device detection happens in the very first page, once you’re on disney.go.com you are going to get the desktop site with flash, javascript, etc

  • the real mobile site is at m.disney.go.com, once you’re there you will always get the mobile site

  • the iPhone seems to go to the desktop site, but then gets a special layout

  • no matter what headers you set, if you go to .mobi you will get the mobile site

  • the desktop site is served with a Windows server, so I assume developed in .NET

  • the mobile site is served from a Coyote linux, using Tomcat

  • if you only provide the User-Agent string of a mobile device you will NOT be detected as mobile

  • in order to be detected as mobile you will have to provide a valid UAProf URI, and that’s actually enough by itself

  • the iPhone on disney.mobi gets the same presentation as all other devices, i.e. not the same as if you visited disney.com in the first place

There is definitely some broken flow here, although I think the mobile experience is very good and the detection at the root is a very good move. Now, will someone from Disney tell me how many visits they get from mobile compared to desktop? I’d love that. 😉

Some nerdy stuff if you wanted to replicate my tests.

Basic request (and then follow the redirects):

curl -D – http://disney.com

Basic request with mobile User-Agent:

curl -D – -A “Mozilla/4.0 SonyEricssonV640iv/R1EA Browser/NetFront/3.4 Profile/MIDP-2.1 Configuration/CLDC-1.1” http://disney.com/

And with UAProf:

curl -D – -A “Mozilla/4.0 SonyEricssonV640iv/R1EA Browser/NetFront/3.4 Profile/MIDP-2.1 Configuration/CLDC-1.1” -H ‘X-Wap-Profile: http://wap.sonyericsson.com/UAprof/V640iR101-3G.xml’ http://disney.com/

And just in case, with iPhone User-Agent (and no UAProf as it doesn’t serve it):

curl -D – -A “Mozilla/5.0 (iPhone; U; CPU iPhone OS 2_2_1 like Mac OS X; it-it) AppleWebKit/525.18.1 (KHTML, like Gecko) Version/3.1.1 Mobile/5H11 Safari/525.20” http://disney.com

Even more nerdy note, the desktop site sends the following header:

X-Cnection: close

According to a quick search some proxy or gateway in the middle.

Speaking of crappy User-Agent strings

I am normally a fan of Opera Mini and I use it quite often on my V640i, but yesterday we stumbled on a very wierd string.

A Samsung SGH-E740 (on Device Anywhere, so you can probably try it yourself) has Opera Mini installed and the User-Agent string is Opera/8.01 (J2ME/MIDP; Opera Mini/1.1.7621/hifi/tmobile/uk; Motorola V3; en; U; ssr).

Now WHY is that “Motorola V3” string there? Surely this is not a Razr V3, surely Opera Mini aspires to be a better browser than the one pre-installed on the Razr V3, so WHY?

I don’t have a clue, of course.

PS: The X-OperaMini-Phone-Ua header is there and has the original User-Agent string.

Sorry, your HTTP headers are incomplete

You might know by now that over the years I have developed a little fetish for HTTP request headers of mobile devices. At dotMobi this is a common reason to make fun of me, they let me discuss them for a little while and then either all walk away or just point out that I’m the only one who cares. Obviously, this is not helping very much my mental issues, so I’m here telling the world.

We all know how Apple is sometimes evil and how they tell everyone that the iPhone has a “full web browser” and is not mobile and bla, bla, bla. We know this is not really the case, but we certainly don’t want to ruin the only one thing in 10 years that might be making the mobile web take off!
The headers the iPhone sends do not provide a UAProf URL and all iPhones (2G, 3G and who knows next) all send the same User-Agent string where the only difference is the firmware revision. It is good if you want to know if it’s OS 1 or 2, useless if you want to know if it’s 3G, has GPS and so on.

Unfortunately it looks like Android is following the same path and not really helping developers. The G1 was the very first GooglePhone and everyone implemented a G1-specific UI or site. Hopefully more Android-based devices will come in the next few months and we will see much more activity on our websites, now, if this is true, be prepared for a new device detection nightmare. See these two User-Agents that we recorded in DeviceAtlas:

Mozilla/5.0 (Linux; U; Android 1.5; de-de; HTC Magic Build/CRA86) AppleWebKit/528.5+ (KHTML, like Gecko) Version/3.1.2 Mobile Safari/525.20.1
Mozilla/5.0 (Linux; U; Android 1.5; en-gb; HTC Magic Build/CRA71C) AppleWebKit/528.5+ (KHTML, like Gecko) Version/3.1.2 Mobile Safari/525.20.1

You might have guessed they come from an HTC Magic, the device that Vodafone should be releasing any day now. If you know how most device detection algorithms work you will know that they normally just walk the User-Agent string from left to right and try to match it with known strings. Notice how “de-de” and “en-gb” are before the string “HTC Magic” this will either break that search algorithm or make us record all possible combinations of languages.
But don’t panic, YET, the User-Agent string is not the only HTTP header that a browser will send when requesting some content. Since this is a mobile device (it IS a mobile device, right?) you might expect a UAProf URL. Even though UAProf has not solved all our issues, it was still one of the very few things we KNEW every mobile device would provide since probably year 2002 or so. It was not the case for the iPhone and the G1 and you will not be so surprised to discover that it is also missing with the HTC Magic.

I am sure I can thank HTC for this, not sure how much Google is responsible, surely they haven’t done much NOT to make it happen. I guess we’ll have to wait for some other vendor to come up with an Android-based device.

One last bit of my rant is about a NEW header that is added, instead of the Referrer, the browser sends a header called Origin, which is EXACTLY like the referrer, but with a different name! Good idea, isn’t it?

logme.mobi update

I felt like my little site to see a browser’s HTTP request headers, logme.mobi, would at some point die and that is why I have not updated it in months. The reality is much brighter and while I see crawlers visiting every other day (don’t care about those very much and I’m sure they do not care about the service!), I also see mobile devices and strange User-Agents coming every day. This is of course a good sign, a proof that it is useful to some and even if it’s very simple in concept it’s good that it is being used.

For this reason I have spent half an hour today to make a small change that I think can be very useful. Up until yesterday I showed the headers as PHP stores them which was probably OK, but less than perfect as PHP uses its own header names and changes everything to uppercase. Using a simple function (apache_request_headers) I have now changed to the actualy names as received by Apache. This is probably a change that will not make a huge difference to the most, but it’s a valuable improvement to some. I have some other improvement on my mind, but I need to install some software and unfortunately I don’t have time, but expect something soon.

I also removed the officially-dead-for-quite-a-while list of tests. I originally created them thinking of using them as a test suite for browsers and mobile devices and wanted to store results in a database. Eventually and thankfully, dotMobi came in and we developed a nice site for testing that is fully integrated with DeviceAtlas and so now I’m linking that site, if you want to run any tests (see the online docs for TA-DA). Remember that in order to login you will need valid mobiForge/DeviceAtlas credentials; on the up-side, all your tests results are stored with your profile.

New Gmail mobile built on HTML5

I was very pleased to read an article from Alex@Google that describes how they have decided to develop the new gmail mobile web interface.

There are at least two reasons why I liked this article, one is that, as Alex mentions, the team originally developed a J2ME application (that I used quite a bit on my old Sony Ericsson W810i) and then decided they needed a web application to serve slightly different needs (and probably slightly different users). The second reason is that it seems like if you don’t create an application for the App Store you are going to fail, while Alex explains quite a few reasons why developing a web site was better than a native application.

Bottom line is, of course, that users get multiple options and an opportunity to choose what suits them best.

I think the article is well worth 10 minutes of your life, if you still haven’t read it, hurry up to check HTML5 and WebKit pave the way for mobile web applications.

Google crawls and converts WML pages

I knew the Googlebot Mobile visited and stored results for XHTML-MP and WML pages and I assumed they would be used as results to mobile users.

It looks like not only mobile users get “desktop” results, but also the other way around.

I was searching, as usual, for some strange User-Agent string and I found this result:

The converted page looks OK in my browser, of course, but I have no clue what it says! 🙂

Swedish Beers @ MWC 2009

In the last two years, when going to the 3GSM first and MWC last year, I always joined Swedish Beers. While it is a very informal event it is actually very well populated and it’s a very good opportunity to have a good beer, if you like and to meet good guys in the mobile space.

This year I am definitely planning to join the event once again and I look forward to meet lots of mobilists there!

Thanks to Helen for arranging everything, of course!

PS: Yes, dotMobi is a sponsor, but we sponsor because it’s a great event!

WordPress Mobile Plugin by Andy Moore does dirty things

Last week I wanted to take a look at the recently released WordPress 2.7 and of course wanted to give it a go on a mobile. As you might have noticed I’m on Google’s blogger and there is, unfortunately, no mobile version, so a proper mobile plugin would be a big plus for me and a good reason to move away.

I download the tiny zip from the official site, opened the readme.txt file. The file itself did not tell much if not that it would make my blog mobile. Fair enough, I took the php file and copied it in wp-content/plugins. That is all that was apparently required to install and in fact in the admin interface I had a new plugin available, I selected it and enabled on my Mac. The site URL was a local one, of course, wordpress.local (and added to my hosts file to access it).

Looking for some further information I went back to the official site (that I won’t link) and notice that in homepage there’s a nice link mentioning that my wordpress.local is the latest site which installed the plugin! A bit surprised I opened the small PHP file and noticed that there are two calls, one on plugin activation and one on deactivation. Being PHP it was not so hard to find what it does and I was very disappointed with the discovery. On both events the plugin calls a remote API AND send an e-mail to Andy Moore mentioning the site name, URL, and the admin’t email, YES, YOUR E-MAIL address.
In my case, the SMTP server was down, so the e-mail did not get delivered, but to my great disappointment the API was reached and that is why my local install was mentioned on the website.

Nothing wrong has happened, in fact I’m pretty sure if Andy wanted to contact me he would definitely find a way, but it was very disturbing to discover that this happens without letting the user know. If I had been a bit smarter I would have looked at the code and I would have discovered it, the problem is that the average wordpress user will not bother to look at the code and will probably not even be capable of understanding what’s going on. We are not all developers and my impression is that Andy is relying exactly on the inability of his users to disable the feature. The software is GPL and everyone is welcome to look at and change it and that is exactly why I did not even bother to look at it, I took it granted that nothing bad would happen.

In case you wanted to disable this functionality (but it’s probably too late now, I admit), what you should do is open wordpress-mobile.php, find the function wordpress_mobile_plugin_activate (it’s at line 1664 in version 1.3), go to the first line of the function and just add the following line:
return true;

Do the same for the function wordpress_mobile_plugin_deactivate (line 1673 in version 1.3).

Andy has definitely spent time to get this plugin working and maintaing it and I think it’s perfectly fair for him to ask for money and ask his users to provide usage details, but asking and taking are different operations, I my opinion.

Vodafone V640i with Netfront, one of the worst mobile browsing experiences

This summer I had the opportunity of getting a new phone for free (or for a small fee on some models) included in a new contract that dotMobi got for me. I thought it would be a good idea to get my hands on one of those phones that every normal customer get with their contract. Having loved my Sony Ericsson W810i I thought it would be wise to get another Sony Ericsson, so I chose the V640i, Vodafone exclusive. The phone is not exactly the lowest end on the market in fact it features a very good media player (including a good music player and playlist manager), bluetooth, stereo headphones included and a few more things. The camera is really cheap, very far from the now really old one included in my W810i. Per my personal tastes, the keys are a bit small and the green and red ones are too close to the accesskeys, but overall (and considering it was free) it’s a good phone and does all you would expect from a basic phone and more.

What I was most curious, of course, is the browser. The experience has been very poor and even if the device has the right accesskey configured to connect to Vodafone Live, the NetFront browser is really really bad. While the browser is capable of rendering WML, XHTML Basic and some basic HTML pages, I really do not see any reason why a user or customer would ever use this more than twice. I have a few screenshots that will prove my points, but let’s start saying that while I was browsing I often had no clue if the browser was retrieving content or not, in fact it was often stuck “loading”, but stopping and reloading the page would make it appear instantly, the browser really seems stalled from time to time. Initially I was in doubt if it was the phone, the network or the browser, but once I installed Opera Mini I was sure it was the browser.

My first test site is Metajam, a site about movies, TV shows and Music. I recommend that you click and see how the page is normally presented so that you get an idea. (Note: the side design has slightly changed since these screenshots were taken)

This is the first thing that I am presented when visiting the site:

At first I thought the site did not work on the V640i.

Clicking on the options accesskey (the left one), I see this item-list where the first item is “accept” (sorry for the poor quality):

I assume this does not tell anything to anyone, but to me, it immediately reminded of the form wizards in some WAP 1 browsers, so I clicked it (and wished I was not accepting something strange such as transferring 1 million dollars to a Nigerian bank).

This is what I saw once clicked, the real page body. Notice how each link is on an individual line, SO 1999:

Clicking again “options” and “accept” I get this very helpful page that I have no clue where it comes from and how it would be of any use:

Later I was looking for a mobie and did not want to browse movies by title, I rather wanted to search by keyword, but I could not find a search menu even though I knew on the iPhone version I had seen one! I took me quite a few pages before I thought about the “options” key. Again, “options” and “accept” was the unexpected solution. Here is another form wizard that you can see only using the “accept” menu. Clicking on the two square brakets opens a blank page where you can type the text and confirm and then be back here, not even automatically go to search.

Again very poor usability!

Enough with forms, let’s see how unordered lists look like. This time I am visiting ta-da.mobi, the test site of DeviceAtlas. Once logged in I go in the “Display” section to test the display size. As you can see the section has two tests and there is a bulleted list:

This is another page in TA-DA, with another bulleted list. The markup is EXACTLY the same, but we have no clue why the browser is displaying a broken image instead of the bullet:

If you are curious, this is the markup of the first page (the one that works):


And this is the code of the second page:

Again, the user will have no clue what’s happening and will think that this is a poor site that doesn’t even have the images in the right paths.

Browsing experience is really poor and some features such as landscape browsing are completely useless and the usability is even worse than with the normal browser. Developer efforts would have been better spent on making the browser work properly. 😉

On the bright side, Opera Mini 4.1 (I have not tested 4.2, yet) works like a charm, it is fast and reliable and both mobile and desktop sites work as expected. A must have for all V640i owners (and maybe all NetFront browser owners).