iPhone 3G Impressions and Development

apple.gifI was lucky enough to grab an O2 3G iPhone online last Monday when they were briefly offering pre-ordering online. Here are my first impressions together with a few development implications. I’ll concentrate on the new features.

I am surprised the iPhone App store is so populated. OK, there are lots of e-books but there are also many real applications. Neither Windows Mobile nor S60 achieved this number of apps in such a short space of time. There’s one catch – you can’t download the free apps until you have added a credit card to your Apple ID based account. If you don’t, you just get the unhelpful error "Authorization Failed, Please connect to iTunes".

One observation I have is that it’s going to be difficult to discover apps on the phone – especially when there becomes thousands rather than hundreds. The iPhone’s simple but clear standard UI only allows viewing of a few on the screen at any one time. In time, expect they will have to create a full screen application (Apple call it an immersive application) to allow more applications to be discovered at any one time. Apps can also be discovered and downloaded on the Mac so it’s not too great a problem.

GPS? At first I thought it wasn’t working. I took it inside, outside and I just got the busy icon in the maps app. While writing this, I left it searching and it’s got a fix… but took 15 minutes! It’s currently accurate to about a 3m and it successfully tracks me within a few seconds as I move around. I am not sure yet if the long lock time is a one off thing or not.

Despite 3G, the iTunes application still insists on a WiFi connection to download. This is a pain for me as I have now standardised on power line (IP over mains) networking for reliability. My WiFi access point has now come out of retirement.

It’s interesting that people have been approaching me with cross platform native projects that include iPhone but not Android development. I think this is mainly because the iPhone is a real device and it’s already being offered by 22 network operators. Android still has a long way to go in this respect.

In what little time I have (I am very busy at the moment), I am getting up to speed with iPhone development. I can’t talk much about the SDK because you sign away such rights when you agree to the SDK agreement. However, generally, I am seeing how you use the iPhone screen assets (table Views, toolbars, tabbars, alerts etc) being more of an issue than learning the Cocoa/Objective-C idioms.

It’s interesting that of the initial set of released applications, very few actually need to be native. However, it all depends on how good your connection is at any one time – online apps obviously need to connect.

Another interesting observation is that Apple is allowing applications to be free but ad-funded. This is good news as the current trend seems to be for applications of this type.

Netsize Guide

netsize.gifNetsize have just published their (free to download) Netsize guide

"The Netsize Guide is a comprehensive annual industry report, documenting the state of the global mobile content and services market. Drawing on in-depth research and exclusive interviews with industry executives and opinion-makers, it provides professionals an essential snapshot of the developments that have taken place and the trends that matter."




The Guide is huge and contains so much information that it’s certain to be useful for anyone working or anticipating working in mobile. I was particularly interested in Rich Miner’s (VP Wireless Technology at Google) views on mobile advertising and his comments on the possibility of carriers having phones open to 3rd party extension. The Guide also includes sections on mobile enterprise and machine to machine – areas of opportunity that are often overlooked. The Guide also includes detailed statistics on a country-by-country basis.

Nomadic Applications

moto.gifThere’s a thought provoking article at The Register on how Motorola think developers should be turning to mobile to capture the large market. The reality is probably somewhat different. Fragmentation of Java (i.e. differences in implementation) across mobile devices is currently the largest barrier to taking advantage of this lucrative market.

However, one comment made me think…

"So one of the key targets developers should hold in mind is "nomadic" – any application that can benefit from being performed where the user is located, rather than where the tools are normally located (such as the office desk and desktop PC, or the hi-fi system at home) is going to stand a good chance of success."

The idea is that with more and mobile devices able to connect affordably (and easily) to the Internet via 3G and WiFi, a new class of applications become more viable. Think beyond email, instant messaging or uploading photos. Think in terms of remote control or remote monitoring and many interesting applications become possible. Here are some examples…

  • Programme your TV set-top box from your phone when you forget to record a TV programme
  • Automatically and silently sync your work PC appointments with your phone
  • Remote monitor the location of children, animals or even valuable items

Some of these things have already been implemented – but not in any affordable or accessible way for the mass market. Another key thing here is that they should be ‘must have’ applications which are indispensable to the user in a particular situation – i.e. Waiting to get back to a work or home PC simply isn’t an option. These are the kinds of things that will cause people to buy a particular phone (or use a particular network operator! i.e. deep Personalisation) to gain access to the service.

Mobile Data Rip-Off?

openhand.gifThere’s a thought provoking piece on The Register about how OpenHand, an email software provider, has found it hard to sell its service through network operators.

OpenHand says that network operators are not taking up their service because it costs less (than say BlackBerry RIM) which translates to less profit for the network operator. Furthermore, the service is more open which doesn’t suit the network operators’ need for services that lock the user in beyond the initial length of their contracts.

Looking beyond OpenHand’s ‘Let’s create some PR via a feel sorry for me story’, this is a problem I have seen with lots of companies I have worked with. Network operators are not very altruistic. They want payback now. A mobile solution must be a good financial proposition for the network operator if you want them to cooperate.  

How does this affect new mobile applications? Well, if you want a large slice of the market you currently have to partner with the operators. This means designing (or rearranging) your product or service so that it looks attractive to a network operator…. both financially and in creating facilities such as customisation that lock in users.

In the longer term I expect fixed price ‘all you can eat’ and ‘WiFi inclusive’ data plans to affect the balance between network operators and third party providers. Released from the high (and ambiguous) cost of data, companies such as OpenHand will prosper.

Orange Mobile Data Forum

orange_logo.gifThis week I attended the Orange Mobile Data Forum in London. It’s a popular event with over 1300 registrations across the six UK locations in October. Orange gave a compelling presentation on why enterprises should consider Orange as their sole network operator. There was also a very interesting technical presentation.

Some snippets…

  • 2005 is seen by Accenture and Orange as the key year for significant uptake of wireless in the enterprise.
  • A recent study of 200 CEOs revealed that they saw wireless as the most important technology.
  • In 2005, Orange will be introducing a wireless convergence product called ‘Business Anywhere’ that will greatly simplify connection to 3G, Wi-Fi, lan and dial up based on enterprise defined rules. The worker will just select ‘Connect’ and the appropriate service will be used based on availability. Furthermore, Orange is partnering with some Wi-Fi hotspot providers so that there will only be one bill. (This won’t exclude use of others which would obviously be billed separately).
  • In time, it will be possible to perform transparent handover between Wi-Fi, lan, 3G and dial up based on availability.

Some technicals:

  • The Orange 2G network currently has about 13000 sites and covers 99% of the population (88% of the landmass). GPRS supports up to 30kbits/sec down and 10kbits/sec up with typically 750ms latency.
  • The 3G network currently has about 4000 sites and covers 70% of the population (80% by end of 2005). It supports up to 384kbits/sec down and 128kbits/sec up. The latency is typically 250ms. [Note that these are theoretical values based on no contention (sharing). A recent AirCom study of indoor 3G access access showed Orange and Vodafone to be more typical 178kbits/sec-198kbits/sec down and 43kbits/sec-44 kbits/sec up]. These speeds will increase to 384kbits/sec up and down in 2008. 3G coverage should reach population saturation in 2008.
  • HSDPA, in trials from H1 2006 should support 1 to 2 Mbit/sec and 120ms latency.

Obviously, Orange sees a lot of added value in providing services as well as connectivity. However, unlike the consumer space, they are being more sensible and not trying to own everything. They realise partnerships are the key. Orange can’t provide everything. This includes partnerships with 3rd party service providers and partnerships with service companies as evidenced by the presence and presentation by Louise Lackenby, a partner at Accenture.

It was interesting that Orange was selling it’s ‘Business Anywhere’ service as a liberating service. Some people might see full time voice and data connectivity as an intrusive and oppressive thing. Orange argues that it allows greater use to be made of dead time, for example on trains and in airports, freeing up personal time later in the day when you don’t otherwise have to catch up.

VOIP currently isn’t feasible over GPRS and 3G due to high latency (round trip times). VOIP will become more and more possible as the mobile networks improve and latency decreases. In fact, large companies sometimes use a leased line connected direct to mobile networks for increased security (data not going over internet) and reliability. In time, these may be used for voice as well as data traffic.

As an aside, at the conference I also came across an interesting WML and JAVA client side wireless caching system by Lan2Lan for the BlackBerry.

Update: 13 October 2005: Orange now have the presentations for this event online. Unfortunately, they have excluded the interesting technical session.

3 Still Limited

three.gifI have avoided UK network operator 3 since their inception due to the fact that they don’t allow access to Internet content outside their own network. This "Walled Garden" has meant that many companies also haven’t even gone anywhere near 3.

As from the middle of September 2005, 3 allow access to mobile content via their ‘Mobile Web‘ service. What more, it’s only £2.50/month for up to 5Mb. It had to happen. 3 no longer has the advantage of being the only UK 3G operator and has arguably squandered the opportunity.

However, there’s a catch. It only includes sites that 3 has deemed are a high enough quality and are suitable for mobile. Sites will be added over time as people request them and they get reviewed.

So where do I go to add a new mobile web site? How long will it take to get reviewed? Where does this leave innovative http (non-browser based) applications that send and receive pictures and video? How does 3 expect to continuously review thousands of sites to ensure they remain of an acceptably high standard? How much does it cost when you exceed the 5MB? I have a feeling 3 still don’t get it.

Data Costs

As a developer of mobile applications I currently see data costs as a serious limitation to the short term growth of innovative mobile applications and services…

We are told 3G is all about receiving (and sending) music, images and video to and from our phones. Unfortunately, it’s far too expensive. This is not only deterring people from using the few services that currently exist but is also making new services a ‘non-starter’ for developers considering developing 3G applications. Look at Virgin’s radio service over 3G – it’s a great idea but it costs more to listen to it for a few hours than buying, say, the latest and greatest portable MP3/radio! There’s absolutely no benefit of listening to radio over 3G for the excessive extra cost.

A survey was commissioned last year by Nokia, conducted by a NOP World in the UK, US, Germany, France, Denmark, Estonia, South Korea and Greece.

  • 31% said they preferred unlimited-use payments.
  • 9% said they were happy to pay according to the data consumed.
  • 11% said they preferred to pay for a fixed amount of data per month.
  • 24% said they felt there should be no charge extra charge for data services.

The key thing is that people either need to know exactly how much it will cost to access a particular service OR they need to know they can access everything for a fixed cost. Event based billing, where you pay per song or per video, only tends to work when the network operator is in control of the service as they can discount or zero the data usage costs.

If you are a 3rd party service provider you might have to do a deal with each network operator and give them some of your revenue in exchange for zero or reduced data costs to the end user. While this is difficult and time-consuming to set up, I have worked with a few companies who have already achieved this within a limited geographic scope. Other options include designing and producing services that use minimal data or even choosing programmatic SMS and WAP Push where this is appropriate.

Meanwhile, the network operators are currently playing with their 3G offerings in an attempt to trap subscribers. Orange in the UK offer unlimited 3G data – but only for 3 months of an 18 month contract. Vodafone UK offer event based pricing but charge you heavily per Mb when you start accessing services outside of their network. Hence they tempt you in with what they know you want – and later you end up paying heavily to help offset the cost of their expensive 3G licenses.

In the longer term, 3rd party service providers are relying on data tariffs becoming much more affordable as competition sets in.