coupling the OS and the Internet
when i was reading through my rss feeds, i came upon this article about ODF support in OOo. if you have no idea what i’m talking about, the one sentence version is that it allows for you to edit a wiki online inside OO writer. it connects them automagically and changes you make locally gets propagated to the server.
this got me to take a step back and assess the state of web applications today. everyone has been pushing for a more native, application-like feel for web pages. in fact, they’re now web applications. with technologies such as javascript, ajax, comet, and other new hotnesses like 280 slides, webapps feel more fluid than ever. there is still something missing. one problem webapps have is that they don’t integrate with the operating system. this observation comes from the fact that there are all these desktop apps for popular webapps such as twitterrific, tumbleweed, and alert thingy. all these applications are social apps, but the same thing goes for other types as well. your work experience on your computer should be coherent and not segmented. segmentation comes when you have a browser and create a sub-environment inside it. there is a strict barrier between your operating environment verses the browser environment only to be traversed by clipboard text.
there are compelling reasons to host applications on the internet. they run on faster server machines, allowing us to timeshare resources and allow us to perform computing intensive tasks without a powerful local hardware investment. the information available on the internet in the application domain you are operating in is also richer than that of your local computing device. yes you can download information from the internet, but that will be an on-the-spot, download and calculate operation. in the internet application domain, that information should be already there and maybe even semi-reduced into a more useful form. it makes a lot of sense for rich internet web applications and thin clients to exist.
the other problem is getting people to use them. whether it is optimal or not, i dont know, but people have grow accustomed to a certain way of working. they are used to the personal computer, and with it, a personal operating system. the web sub-environment doesnt fulfill that requirement. we need to bridge the gap.
there are two parts of this. actually, there are three parts, but one of them has already been solved (to a certain extent). the solved piece is that of the local computer. this is what people spent most of the 80s and 90s doing. there are operating systems, applications, ways to store data, document formats, and anything you could possibly want right out of that box you buy at the store. the second part is getting your information organized online. currently, documents are scattered about on different domains. you may have your personal likes and dislikes over at myspace, your school papers on google docs, your pictures on flickr, your source code on github, and your videos on vimeo. new services keep coming out that do pretty much the same thing. last year you may have been using myspace, but this year you’re on facebook. new online photo editing sites keep coming out, and there are many online video sites that hope to capture the magic that happened to youtube. after a while, you’ll have your data scattered througout the internet. things like opensocial, atomkeep, and the data portability initiative aims to consolidate your data and keep it organized.
after your online and offline data situation is not an embarrassing mess, the final piece is connecting it together. currently, you use some local apps to do some stuff and then some of it can be done on websites which is not a very compelling user experience. we need to find a solution that cleanly extends your local computer experience by augmenting your local resources with that of the internet. this may be more of a usability and human interaction problem. the technologies seem pretty innocuous. for example, maybe you can provide some contextual actions such as editing a document with google docs and then synchronizing the results back to the local machine. or you can integrate help menus with information from the internet, formatted in a way that resembles local information. all of this requires standards because there are usually more than one party involved in this type of integration. either a product is solely web based (google docs) or solely OS based (MS Office). companies may want to start providing hybrid solutions in specific vertical slices. this doesnt seem so important for things like document editing, because you can create a doc editor in both the OS and the web. they will both have different pros and cons, but it can be done and will achieve pretty much the same effect.
i foresee a new generation of applications that in order to be effective will need to draw upon the native interface and integration provided by the operating system coupled with the data and hardware resources available from the internet.
