I'm in a presentation with Matt Webb who's just been talking about his ideas of web applications in the next five years or so. He's made some very good observations about the way things are going, I'm going to list each of his points and try and expand on them a bit. 1. The next Mac OSX is going to ship with RoR, you will be able to run your web app locally and host it in a heavily customised browser interface, you can allow script in the app to get access to things that are usually protected, such as the file system. Well it's not just OSX with that stuff, IE has had HTA since IE4, .net has the casini web server, which is lightweight and won't conflict with other instances or iis. You also have XAML for building very powerful app interfaces in IE, or XUL for firefox and flex for flash if you want cross platform. (More of flex after Arals presentation coming up next) 2. Massive, Scalable, Affordable Need lots of storage, lots of processor power? Amazon have it all, and your users can just pay per use. Say you want to build a photo management site, if you ask your users for their S3 details you can back all the image storage onto their account and they pay for it. If they want to run face recognition across their entire photo library you can farm that out to EC2. There are many other similar services such as Alexa for indexing and searching. Personally I see these technologies as a way for quickly growing an application, once you have the ad revenue you'd want to host all this stuff yourself, latency will always be a problem, as will the bandwidth bills. 3. Location and context intelligent applications Matts point was that if your laptop knows pretty much where it is, based on wifi network etc, it should be able to change its interface to what you need, depending whether you're at work or home. I think this is something we could start using now, based on a users IP, imagine being able to mark photos or groups in your flickr to never show up when you're at work. You could do this with cookies so people choose per computer, or based on IP so they choose per network connection. 4. Ambient applications not interrupting Matt worked on some really interesting continous partial attention prototypes, someone else mentioned Growl, which looks a lot like the little popups office and other windows applications are starting adopt instead of that bloody blinking taskbar button, a popup, or a noise. 5. Web Pipes I'm sure most of us are familiar with pipes in unix, I'm a windows user but I love my windows ports of all those great unix utils like grep and less. Matt mentioned apps like filtr and scanr, that process images and allow you to send them on. There is also preloadr that gives you loads of image processing online before passing on to flickr. And of course you can pipe your feeds through the likes of feedburner. Daisy chaining these things together allows a none techy user to build some very powerful customisable applications. One point Matt made is that it can be quite difficult to monetise these things when its just one step in a long pipe, they can't all insert an ad! 6. Interactive RSS We're seeing more of those little links down at the bottom of items on feeds, like digg me (look down). Matt thinks we're going to see more of this, such as embedded forms, so when you get a new bug added to your bug tracker, you see it in your reader, choose the developer to assign to, click and its done. Some of the things he talked about have been around for quite a while, and we're going to see a lot more of it in a lot sooner than 5 years.

Labels: barcamp, barcamplondon, barcamplondon06, future, matt webb