Comet is at the intersection of 2 trends. Put the 2 trends together and you have a compelling case for why the time is now right for comet.
Trend 1: The time spent on a single web page is increasing
It doesn’t take a lot to demonstrate this trend: Ajax-y web pages existed before the coining of the term, but only in pockets. Today it is a standard part of every “web2.0″ startup’s buzzword compliance list.
As the use of Ajax increases, so does the time people spend on a page without navigating to a new page increases. It’s been noted many times that Ajax means the death of the ‘page-view’ model of how successful a website is.
Through hours of painstaking research into the growth of Ajax on the web, it’s possible to create the following graph:
Trend 2: Pages are becoming more dynamic
There are a number of features that feed into this trend. The first is that websites are becoming smarter and smarter. We are now able to do things with websites that we didn’t think possible 10 years ago, like office suites and multi-player games.
The second, perhaps more important trend is the growth of the social side of the web. The web used to be somewhere you looked up recipes, now it’s a place where cooks create social networks to comment on what they’re cooking, who they got the recipe from and how it turned out.
It’s easy to see that the half-life of the text on a page is going down. From Facebook news feeds to blog posts and ratings sites, the content of the web is changing faster than ever.
Again, through hours of painstaking research into the half-life of web pages it’s possible to derive the following graph:
The Growth of Comet
It’s easy to see why comet is such a sure bet. On one hand we’re spending longer and longer on a single page, but on the other hand the chance that the page we’re looking at has changed on the server is going up. Clearly we need a simple, asynchronous, low latency way to update pages. Comet is the answer because it lets you do just that.
As final proof, through hours of painstaking cutting and pasting of the graphs above, we can create the ultimate executive proof of the growing need for comet:
Clearly the crossover point for any web page will be different. For chat sites like Meebo, the crossover happened years ago. For office suites that allow parallel editing, the crossover has also already happened. Sites covering live events are next in accepting comet, and it won’t be long before news feeds in social networks start dynamically updating too.
Comet isn’t going to evolve to be used on every web page on the Internet, but it is going to find its way into places that we’ve not thought of yet. 15 years of plain HTTP request/response has taught us to think in terms of static pages that don’t change by themselves too much. Comet lets us treat the web as a network of people interacting more naturally than they can behind a request/response barrier.