Unintentionally Blank

Phil Nash on the Internet, Web Standards and Accessibility

Future Of Web Apps: Day 2 In Detail

Oct 09, 2007

by Phil Nash

Future of Web Apps

Last week saw the Future of Web Apps conference come to London. I was there for day 2, listening to presentations from Facebook, Adobe and Yahoo! among others. With any luck, I also figured out what the future of web apps is from just one day, here is my impression of the presentations and exposition after a few days to think about it.

From The Top

First up was Paul Graham of Y Combinator talking about the future of web startups. I missed most of the talk, sadly, but I think the gist of it was that the future is bright for startups with better funding opportunities and better infrastructure in the web sphere. This seems directly in contention with the view of things at Read/Write Web, as they continue to consider the digestion phase for Web 2.0. While there may be a bit of thought on behalf of some of the larger competitors in the web, there will be greater and greater influx of smaller companies with new and interesting services.

I'll stop there as I only caught the end of Paul's talk, however I will be reading the essay derived from his keynote, also titled The Future Of Web Startups.

Adobe AIR

Leah Culver of Pownce was talking next on the developer stage, but, as I am a developer first, I headed to the back of the conference room to the Adobe stand to see what they had to say on their not-out-of-beta-yet technology, AIR.

AIR allows you to develop desktop applications using web technology. AIR combines HTML, Javascript, CSS, Flash and Flex to produce applications that work offline and connect seamlessly to the web while online. It sounds like I am part of the Adobe marketing team when I speak like this, but I am actually excited by the prospect. The emphasis of the conference is about web apps, but people have been increasingly after taking them offline (even Firefox 3 is supposed to support offline web apps). But what makes AIR so simple is that it uses and extends current technology and standards (as far as HTML, CSS and Javascript anyway), technology that we already know how to use and program with, to extend our online web apps and take them offline. It will also be platform independent, there are PC and Mac versions of AIR at the moment and a Linux version is promised by the time it is fully released or shortly after. There is even a chance of a mobile version later down the line.

The AIR presentation showed us how to create a simple AIR program (everyone loves hello world) as well as slightly more involved examples including a simple real time HTML editor built in Flex and a bunch of bouncing red balls built in Flash which showed off the optional transparency of an AIR window.

I hope to cover more ideas and simple AIR programs as I try to pick it up myself over the next few weeks. I'll also have to think of a web app to take offline (or even put online and then take off again). In the meantime, have a look at some of the applications already available for inspiration.

Facebook's Platform and Its Vision For The Present

Back to the Developer stage to see Dave Morin talk about the Facebook Platform. While there were some interesting stats (60 billion page views a month, user base growing by 3% every month and 50% of users returning every day) there was not much about the future. Dave presented the basics of writing Facebook apps, explaining that there exists a Facebook Markup Language (FBML), a query language (FQL) and now a Javascript subset (FBJS), but he didn't go into any depth which I was slightly disappointed about.

The future seemed not to matter though, while Facebook are providing a platform for web apps they looked like they were just sitting back and waiting for developers to come to them. There were a number of slides showing where on a page Facebook would place ads and were the app developer could put theirs, so it seems that monetising is of great importance and Dave also talked about optimising each type of page and area of a members's profile.

Questions at the end of the session brought up worries about security of users' information and worries from developers that big companies could clone their apps and leverage their market to steal users from the small developer. Dave seemed a little unseated by this and could only answer that Facebook will keep innovating and finding ways to help all developers.

All I could think about during this presentation was that however open the Facebook Platform is, when you're within the confines of Facebook, I would be watching out for Google's announcement, apparently on November 5th, which could open up all their services and mount a serious competitor to Facebook.

Short On Cycles, Long On Storage

I was about to leave the Developer stage when Simon Wardley started talking about ducks; I had to stay. He is waiting to upload the video of his amazing 300 slide keynote regarding the economic cycles the web is going through. While I would love to try and explain everything that Simon got through, I implore you instead to wait for the video and watch it.

In summary though, I think Simon's point was that as parts of the industry mature we should see more and more service providers (of hardware, software or applications, think of cheap, outsourced storage like Amazon S3) which will enable creators of future apps to stop worrying about the infrastructure and pay attention to the app. This will also allow for more apps to come onto the scene, backing up Paul Graham's points earlier. Finally, these cycles will continue and while there are some services emerging now, more will on top of the current ones. I thought this was alluding to APIs, where we are now starting to get apps built on hardware and software services from third parties, the future will include more and more apps built on apps.

The Practical Semantic Web

After a break for lunch we reconvened in the Developer stage to hear from John Aizen and Eran Shir from Dapper. I was really looking forward to hearing what they thought of the semantic web and I was happy to hear a few more details from Dapper's perspective on what Read/Write Web recently pondered regarding the top down semantic web.

Dapper itself is a tool with which you can create your own API for any site you want. All you need to do is select elements of a page and Dapper will create a feed based on your labeling of them (see their demo for more detail). I was pleased, also, to hear that Dapper is more than just a screen scraper, it employs more stable analysis of pages and can even stand a small redesign, though major restructuring of pages would impact the quality of "Dapps".

The exciting thing about the top down semantic web is that it is driven by those who want to use it, not by companies who can see no short term gains in supplying it. This is already working with Dapper, and can only continue to improve as the community grows.

APIs All Around

Next up was Matt Biddulph, creator of Dopplr a social network for frequent travelers. He gave an engaging talk on using and providing APIs for your users. He even made the point that Dopplr could be a service that users never visited, journeys could be picked up from online calendars and posted to Facebook, with no work done by the user. This idea of the user owning their data to do with what they wanted is a noble one and one that is possible with some of these emerging technologies.

APIs were part of the services Matt was talking about, but as important was user identity. If various services are willing to give you control of your data to use with other services, there has to be some way of knowing who you are and exchanging that data. We are currently at the stage where we give up our username and password to certain services to see whether our email contacts are using the service. Giving away a password is so insecure and there should be better ways of exchanging data. Thankfully, that day the oAuth 1.0 final draft was released. This is an open service that will be give users access to their data on a service without handing over passwords.

Matt also covered widgets, such as Facebook apps and badges for personal sites as ways that users could use their data. Finally he went on to utility computing, covering a few of the ideas that Simon Wardley mentioned earlier.

Services like Dopplr do seem to be emerging as the next generation, willing to be a part of the internet, not the centre of it.

Being The Middleman

The final presentation I saw was from Tom Coates, on a Yahoo! project called FireEagle. This service hasn't even reached alpha testing yet, but the possibilities were very interesting. FireEagle is billed as a "way to share your location with friends or with other websites and services!" Essentially FireEagle will be able to receive data on your location from any number of services organise the data and provide it to other services of your choice.

The important factor was that FireEagle is being built on open APIs so any service will be able to transmit to or receive data from FireEagle. The possibilities of a service like this are limitless, once various security questions are cleared up (as I said, it's still pre-testing), but imagine being able to get suggestions or information about anything within your proximity, wherever you are. And, because the possibilities are endless, if a service doesn't exist to do what you want, you can create it based on the open APIs.

In Conclusion

So what is the future of web apps?

From what I saw on day 2 of the conference, it seems to me that the future lies in cross breeding services, building on top of each other and meshing the internet together. You will be able to connect your many online profiles and use all the data that each service provides to create both an online picture of yourself as well as services that will help you go about your day. The future lies in open standards, APIs and oAuth. AIR is available for those times when you can't be online and for seamlessly integrating the desktop with the internet, also through APIs (if you can't build yourself an AIR app for your web app, then why not let someone else?). Completing the picture are services like Dapper that are creating APIs for apps whether they like it or not.

The future of web apps is in openness and working together. Now, all I need to do is come up with my own killer app so I can join the party.

Unintentionally Blank is Phil Nash's thoughts on web development from 2006-2008. Any code or opinions may be out of date.