Mobile search is poor - it produces inconsistent results and can often fail to tell the difference between mobile and fixed sites.  This article details the failings...

Let's start with a few of definitions.  The Mobile Internet is the collection of sites which have been specially built/formatted for small screen devices/phones.  The Fixed Internet is the Internet as viewed on large PC screens.  Mobile Search is the attempt to search for content on the whole (fixed and mobile) Internet where the results can be displayed and formatted on small screens (examples are m.google.com and m.yahoo.com) .   The problems for the mobile search engines are that Mobile Internet is still very small so the search engines include (actually they prioritise) content from the Fixed Internet and then display on the phones using a transcoder.  A transcoder takes fixed Internet content and reformats it for the small screen.  Transcoders have what I believe is the impossible job of taking a web page and putting into a long thin column on a small screen and then paginating it if it's too long.  The reason why I believe this job is impossible is because the fixed internet consists of a vaste array of web site formats and there is no way anything can know how to cut up a page and put it neatly in a long thin band on a phone screen in a user friendly way.  Its true that good transcoders can make a reasonable job of 2/3rd of the Internet.  For example 3 column web-site can often display OK on a transcoder, although because the left hand column of web-site is often the navigation this does cause transcoders an issue - some display this navigation at the top of the page and the user has to scroll all the way through it before he can read the content he really wants to.  Others more successfully hide the navigation and take the user direct to the content and then allow the user to open up the navigation later.  The better transcoders also realise their impossible tasks and do not transcode if the device has the capability of displaying a web page (eg for a Series60 running the Safari browser devices or the iPhone).  The good search engines can tell the difference between a mobile site and a web page and not transcode

Now let's get into specifics how do Google and Yahoo mobile search engines compare on these things.  Please be aware this is the siutation in March 2008 and because of the fact that nothing really worksperfectly what is actually happens is changing daily.  So if you read this more than a month after March 2008 be aware it may not be the same.

First both Google and Yahoo present web and mobile web results separately, although Google does put mobile results marked with a little phone symbol in the web results when there is a significant result and real mobile web content is an extra click away on Google.

Finally, I think the main problems result in there not being enough Mobile Internet content.  For me transcoders are an temporary technology in place because there is not enough Mobile content.  Long term webmasters will either create mobile specific sites or use tools such as Mippin to quickly convert them.  Once there is significant Mobile Internet content in place search engines need to start prioritising mobile content over the poorly performing transcoded content.

As I write this the situation changes daily so I hope in a few months I will be blogging a slightly better story for Mobile search.

I've just spent most of the day watching all the iPhone videos at Apple developer site.  All very interesting and informative.  Some of the new Safari features such as CCS transformations, CSS animations and the local data SQL support from Javascript (sort of Google Gears like) look really good.  When are we going to get these features in S60 Safari browser?  Please tell me it's soon!!!

But that was not the point of this post.  I watched one particular video called "User Interface Design for iPhone Applications".  This is an excellent video and lists great advice for developers of any mobile application for any platform.  Basically it says when you take a desktop application and port it to the iPhone you should only take a subset of the functionality and create a simplied user interface AND the UI design should be consistent with the iPhone UI design.  All excellent advice which I would say works for any mobile application design on any platform.  However, the more I watched these things the more I thought of the iPhone advert for the Internet shown on US and UK TV.  See the AT&T version here.  This advert has always bothered me.  Basically it shows the iPhone looking at the PC web version of the New York Times (in the US version) and the Guardian in the UK version.  The claim is that the PC Internet is now works unaltered on the mobile phone.  Any yet this is totally inconsistent with the advice Apple are giving application developers who are telling them to change adapt the content and functionality for the iPhone.

Has Steve Jobs noticed the inconsistency between his marketing message to the consumers and the technical advice to developers?  Web applications are still applications after all and also should be adapted to the phone.  And indeed, this has happened.  Facebook wrote a very good web application for the iPhone.  So why Mr Jobs are you telling consumers and developers totally different stories?  Well, infortunately the consumer is being told an incomplete story yet again which will rebound on the the poor software industry yet again.  So please My Jobs (you love consistency in your user interface), please be consistent in your messages.

I was very interested to read Michael Mace's latest blog about the death of mobile apps, and Carlo Longino's similar blog also taking about the death of mobile native apps. 

Michael Mace shows this graph on his blog

We had this debate a year ago in our company and totally changed direction because of it - basically we stopped producing mobile apps a year ago and started producing a web only solution.  The concern was that we would not be able to create such a good user experience with the Web as we had achieved with our Mobile apps - afterall we had won several user experience awards for our applications.  This fear has not materialised; we put considerable effort into producing an excellent web experience.  Where we've really excelled is in our development cycles.  We've gone from a 3 month development cycle producing 14 different application variants (12 Java and 2 Symbian native variants) to a 3 week development cycle - we actually have more user experience variation than we used to coping with difference screen sizes and device capabilities but the development is so much easier and so much faster. 

Admittedly our application is ideal for rendering in a browser, but other application developers should also consider a move to web experience too.  Afterall the same happened with the fixed internet in the 90s.  In the beginnings of the Internet many applications were produced; my first experience with Internet shopping with Tesco was using an application which required a huge installation including a local database install.  It worked fine for the early adopter (me), but when they wanted to make it a consumer experience they came up with a brick wall - it was called installation (too many things could go wrong and did).  The same is true of mobile applications today - installation is too big a hurdle for the average consumer.... it needs to be just a click away.

The mobile browser is getting better day by day and can now do many of those things only applications could do in the past - its only going to continue getting easier in the browser.