Archive for February, 2013


User Agent (instead of feature) detection

I previously stated on this blog, and in real life, that there are sites that don’t provide the best experience with my Symbian device just because they are using browser’s User Agent string instead of detecting capabilites.

Just for facts’ sake, the User Agent string of my Nokia Symbian phone’s browser is:

Mozilla/5.0 (Symbian/3; Series60/5.5 Nokia808PureView/113.010.1507; Profile/MIDP-2.1 Configuration/CLDC-1.1 ) 
AppleWebKit/535.1 (KHTML, like Gecko) NokiaBrowser/ Mobile Safari/535.1

Apple’s iOS5’s is:

Mozilla/5.0 (iPhone; CPU iPhone OS 5_0 like Mac OS X) 
AppleWebKit/534.46 (KHTML, like Gecko) Version/5.1 Mobile/9A334 Safari/7534.48.3

and Apple’s iOS6’s is:

Mozilla/5.0 (iPhone; CPU iPhone OS 6_0 like Mac OS X) 
AppleWebKit/536.26 (KHTML, like Gecko) Version/6.0 Mobile/10A5376e Safari/8536.25>

First, the reason I listed al these is so that you can see that the webkit build used on Symbian FP2 is actually a version between iOS5 and iOS6’s webkit version. This actually corresponds to each OS’s release date, so no question marks here.

Now, to the more interesting subject of sites:

There is a common belief, along web programmers since the Netscape 4 vs Internet Explorer (4?) wars, that one should try to code a site with conditions on browser’s features, as opposed to conditionals on browser’s name.

Nowadays, this ‘feature detection’ is what allows you to browse your banks’ site with Chrome, IE, Firefox, Camino, Galeon, lynx – you name it. We have passed waaay ahead from the times when ‘Internet Explorer is required’ was shown on public sites.

However, the same does not apply to mobile site, it seems. It may be the case that the mobile devices’ features cannot be detected in a simple manner, so there sure is some browser User Agent string sniffing around.

As a proof, I happen to have a browser named ‘AnyBrowser’ installed on my Symbian, which’s unique selling point was that it surfs the web under an Android browser User Agent string. Other than that, is a simple QWebkit webview, with a location bar, history, tiling & scrolling and other browser chrome (It was an early 0.4.x version and I don’t find it anymore, so excuse me if you cannot repeat my experiment).

So, browsing the web, with the same webkit component (as shipped by the Qt project), under different names – here are the results:

Picasa Web

On Symbian browser vs Symbian AnyBrowser: by default, you are not even presented with a whimsy 360p image



Facebook login

On Symbian browser vs Symbian AnyBrowser: seems to only be CSS until..



Facebook timeline

On Symbian browser vs Symbian AnyBrowser: Is this actually the same site?



Google Mail

On Symbian browser vs Symbian AnyBrowser: Again, google is by far the worst offender

(Picture got removed for privacy reasons, will update it). Actually, forget about it! AnyBrowser can show Google Maps, while the Symbian stock browser only gets redirected away!



On Symbian browser vs Symbian AnyBrowser: notice the use of custom fonts




On Symbian browser vs Symbian AnyBrowser: search and some asset dimensions are better, but not a whole new site



So really, I cannot stress this more:

All mobile platforms (except Windows) have a webkit build. Differentiating between them when doing a website is like differentiating between Mozilla and Firefox!

The worst offender, is, again, google, followed by facebook. – at least in my examples. Google does not even have APIs for its services (think about dropping exchange support for gmail/calendar, then no API for google plus) or if they have, they are not as popular as they think since nobody does 3rd party apps using google services (and lives on).

Facebook at least has 3rd party applications, so it is useful on Symbian.

update with diclaimer: I am not claiming that those sites work perfectly with just switching the user agent string to an Android one – some testing still needs to be conducted by the site’s owner. But that’s the problem, they don’t do this and preffer to hardcode for a subset of webkit engines which may have – as it’s the case with Apple and Google – some vendor specific features.


Public Transportation comes to Bucharest

First of all, let me acknowledge that Public Transport from the Nokia Maps Suite has been available for some time now.

Also, since about 5 years it was possible for me to already look like a native when walking in new cities, using Nokia Maps’ ‘walk’ mode. That thing worked pretty well by putting my N85 in my pocket after inspecting the overall route, and then turning left and right at each new vibration. (As a rant, today’s phone don’t have that kind of powerful vibrator to actually use buzz-guided walk navigation from your pocket, you would still have to hold your hand on the phone in your jacket..)

Adding public transportation to beeps-and-buzzes guided walk means that you look even more like a native when you’re not staring at maps when at subway or in stations for surface transport.

Now this is the new support for Bucharest I was talking about:


If you’ve already been to Bucharest – or you’re living there, for that matter – probably there is no surprise for you that the support is not quite one of the top: there is estimated journey time (with data from Metrorex and RATB – no ‘Cora’ or ‘Real’ supermarket buses:P) but no actual schedules.

This is not a flaw with the application however. There are no real schedules in reality either. Or if they are, they are not respected, partly due to the kind of traffic we get but mostly for unknown reasons.

Now for planning a journey: I set the starting location as … ‘My location’ and the destination the workplace, so I could get the commute time.

It gives me two options: either take the subway, or mix the subway with terrestrial (tramway) transport.  Both of the options are correctly accompanied by some walk needed to get to it and then to get to the final destination.

What’s interesting is that there’s no walk in the alternative route between the subway and the tramway, which actually corresponds to the reality (the stations are in the same place as much as the subway vs. surface allows).

It says ~50 minutes for the first option, which roughly corresponds to what I have witnessed.

Now on to selecting the first options, we have the three segments detailed in a vertical fashion: 12 mins walk, 26 mins subway, 12 mins walk (left picture).

Unfortunately this does not take into account the time for the actual train to get into station, as stated above, because no one really knows.

Now, if you click each of the segments, you get even more usefulness: starting with the subway trip, you get the list of stations you are reaching in your way from Nicolae Grigorescu to Lujerului (taking the train in Preciziei direction).

(see right image)

If however you click one of the ‘walk’ segments, you get the beeps-and-buzzes walk mode navigation mentioned above from your current position to the next station in your trip.

And of course, if you haven’t been living under a rock for the last 5 or so years, you get the option of voice-guided navigation on every Nokia phone since 2007 (for free since 2009) if you would ever choose to commute by using a car.

In this case, using a car gives you a ~40 minutes estimation if you take traffic into account (Traffic however is the next thing I wish Nokia should make available for Bucharest).

Not pictured in the below image, but these are 13 kms.

So yeah, can’t wait to see how this goes when you’re actually in a foreign country and an unknown city! (Altough I can now still test it in Bucharest since it has become available, because for other destinations I could’ve aready tested if I were visiting, y’know).

What can I say? Keep the updates coming, Nokia!