MWI Team Blog

Dispatches from members of the W3C Mobile Web Initiative Team

Categories: Current state (32) | Developing Countries (15) | Events (20) | Looking forward (13) | News (42) | Technical (33) |

Thin and dumb, or fat and smart? — 26 July 2007

Is it better to be thin and dumb, or fat and smart?

This question has been asked and answered in many different ways over the past years in the mobile computing space; in more classical terms, what is the best optimization between making mobile terminals as thin shells while the most of the processing is done on a server, or having mobile terminals as increasingly powerful pocket computers.

This tension has been showing under many different aspects in the mobile web space; consider for instance the development of so-called "full-Web" browsers on mobile devices (iPhone's safari, Opera mobile, Nokia N60), which usually require fairly powerful smartphones, that goes more or less in the opposite direction of "mini-Web" browsers (e.g. Opera mini) where most of the rendering/processing is done on a server and sent to the thin client pre-digested, or services that propose to proxy your navigation to make it mobile friendly, as Google Mobile or Mowser.

It shows also more generally in the Web space: the phrase "web application" has been used ambiguously to designate applications whose logic reside on a Web server, and for which the interaction is done through a Web browser, as well as to talk about rich client applications that use Web technologies - especially with AJAX and widgets being so hype nowadays.

This tension as I understand it, is built on advantages and inconveniences specific to these approaches:

  • if you put most of the processing on the server, your service can be used on many clients (as they require less capabilities/features on the client), you can deploy new versions quickly and reliably; on the downside, making these applications work offline is still very challenging - challenges that Google Gears is at least trying to tackle
  • if you rely more on the client side, you can get a better integration in the interface, use client-specific API (e.g. to take pictures with a camera-phone, or get your localisation with a GPS module); that's what makes Google Maps much more useful as a standalone application on my Tréo than it is in my Web browsers on that same PDA; but getting these applications updated and maintained is usually a much more involved effort, and they often require being developed in multiple versions and levels to cater to the wide diversity of computing platforms.

Of course, the question is then: how do we get the best of the two worlds?

The goal seems clear: getting the two paths to converge at some point in the future; getting there is difficult, but there are a few intermediate steps that seem pretty important to me:

  • there needs to be some serious work on the offline capabilities of mobile devices - Google Gears may be a step in that direction, but I tend to think that some real investment on making HTTP-caching a platform-level operation (the same way getting a network connection is, for instance) would be a tremendeous help - I have already ranted about it for browsing, but that's even more important for Web applications development; more largely, synchronization protocols will become increasingly important - WebDAV, CalDAV should lead the way, but a generic framework for synchronizing data has not crossed my radar yet
  • currently, most browsers don't have access to low-level client APIs; there has been a lot of work done in the Java world on defining these interfaces (through the JSR process), but they remain invisible to the browsers in most cases, and to their client-side scripting capabilities even more so; the W3C Ubiquituous Web Application Working Group is chartered among other things to look at this problem space, and hopefully will make it possible for Web applications to get better integrated in their hosting devices
  • Web access should become much more pervasive in client-side applications; as of today, on most phones, if you get a Web address in an SMS or in your calendar application, you would be hardpressed to use it directly to launch your local browser
  • the level of interoperability for rich client Web technologies needs to improve dramatically to make it possible to rely on them for serious user interfaces; I'm hopeful that the work made by the W3C Compound Document Formats Working Group will help; it seems also to me that increasing the availabilities of test suites for user agents is primordial to achieve that goal

The way I see it, the distinction between Web applications and local applications will be fading; I think Web technologies can serve as a very strong foundation to power this new generation of applications. There are already some ongoing work on development of an open-source platform for a Web-based desktop - I think this approach is even more important for mobile devices.

by Dominique Hazael-Massieux in Looking forward 1 comment Permalink

Is the difference between OLPC and the Mobile Web only about device ? — 24 July 2007

When talking about providing ICT in Developing Countries, everybody heard about the OLPC initiative, also known as the 100$ laptop. Some people may also be aware of the work the W3C MWI Team is starting in that area, focusing on providing Web access on mobile phones.

Obviously, technically, both approaches are quite different. The former is focusing on laptops, wireless and mesh networks, while the later is based on mobile phones and GSM networks. But are these different solutions competing ?

I don't think so. I'm convinced that laptops have their place on the big picture. It is clear that accessing information, communicating and so on is easier with a full keyboard, and a big screen. It is also easy to provide free wireless/mesh networks. Removing one of the barrier, the price of the device, is an essential step. But there are also few other steps to make: deploying those laptops, deploying networks, training professors before they could train children, having people adopting this device,... On these different barriers, countries are not at the same level of development and the resolution of these blocking factors may be quicker in some countries than others.

However, in my view, this is not comparable to the current situation in mobile telephony, where networks are already here, and devices already in the pocket of hundreds million of people. So I'm convinced that the potential today for a short-term impact on people lives is with the mobile platform and Web technologies, but the laptop solution may be an alternative option in a further future.

That said, OLPC and our approaches should not be compared only at the device level. The point is not on technology, but on improving people lives. So the essential points are on how to reach people, how to take their needs into account, how to enable them to use and adapt the technology to their needs and requirements...

On these factors, this is where i think the major differences are between the two approaches. While OLPC has a top-down approach through governments buying laptops and giving them to their children for free, we have a bottom-up approach: relying on existing infrastructure and involving local actors of the society: NGOs with ground expertise, local entrepreneurs, digital divide experts, researchers from local universities...Building such a community is, for us, the only way of capitalizing existing expertise in different domains, as well as really understanding needs, requirements and blocking factors we have to work on. For now, this is the major activity we are focusing on, through the organization of events like the one in India we did last year, and the opening of W3C offices in Developing Countries (Morocco, India, Southern Africa existing, Brazil, Chile, Thailand and more to come) to have W3C presence locally. We are not going to create any new device, or any new technology, but we believe that Web technologies with mobile phones is the only affordable and scalable way of deploying eServices (business, Government, Health...) in rural areas.

Starting from this point, it is essential to understand what types of adaptation of the technology and the environment (here handsets, browsers, networks...) is needed to transform the potential into reality.

Having a very cheap laptop is good, but it would really help people if it allows them to develop business, to earn money and so on. The point is not really the cost of the device but the difference between the cost and the income (see the great talk of Iqbal Quadir on "The power of mobile phone to end poverty").

At this point, while incredibly successful stories are showing up everyday on the usage of mobile phones (read eg fishermen in India story or how mobile phones leverage access to banking services for the poorest people), i don't see similar evidences of the potential impact of laptops. Some specialists tend to think that OLPC hasn't taken these aspects into account and requires high levels of investment before having checked the usability and the potential of this solution to really improve people lives (see for example a very interesting interview from Ken Banks ).

by Stephane Boyera in Developing Countries 4 comments Permalink

Show me your mobile phone... — 24 July 2007

Riding the trains around Tokyo today, i suddenly realized that there is not a single phone ringing, and that no one speaks loudly in their mobiles -- something quite usual in Europe for example. But everyone has a mobile phone in their hands, from the kid to the grand-mother. And everyone is using it, all very naturally. Either to write an email, to check the next train schedule, to order an item and pay for it, or to find where the restaurant they're heading to is located. All these services are available, and in an intuitive way. Hey! another example: Mike Smith just sent an irc message to say that he'll be late for a call. And he is in a train!... But what i like most is to look at these colorful (yes, they have red ones!) mobile phones and the multiple funny looking mobile charms hanging on them. "Show me your mobile phone and I'll tell you who you are"...
by Marie-Claire Forgue in Current state Permalink

<< Previous Page :: Next Page >>

Contacts: Dominique Hazael-Massieux