[SamAtt]

This article basically uses the graph as a jumping off point to make the same old 'apps will take over the world' argument we've been hearing since the iPhone opened its SDK (it seems like people have completely forgotten the evolution of the Desktop PC and are intent on reliving those same mistakes with the smartphone)

The problem I see with the graph is it's based on percentage of total traffic but it doesn't take into account a huge increase in bandwidth and internet users. Put it in perspective and you realize the web's "peak" according to this graph was when most people were using 56k modems (2000 which was about a year after V.90 modems started hitting the street).

There's also an efficiency factor. Ajax type technologies have dramatically shrunk the size of the data chunks being passed back and forth. Back in 2000 most were still resending the whole page on every refresh.

[BerislavLopac]

"Put it in perspective and you realize the web's "peak" according to this graph was when most people were using 56k modems."

I think this is precisely the point: traditional, browser-based Web (text-based, asymmetrical...) is designed to work perfectly with low-bandwidth connections. With the advent of broadband, it becomes obsolete and is to be superseded by modern technologies which better utilize this type of connections.

Of course, Web is not dead by any means; but Wired is in the business of selling tech-related sensations.

[pierrebombay]

Agreed. Aren't these apps just an extension of the web anyway? If it's sent over HTTP and is accessible to virtually anyone who had access to the "normal" web, then isn't that the web.

I know many people are excluded because they don't have iOS running on the machine they use for the web, but not everyone had the means to surf the HTML/Flash-based web on Opera's web browser a while ago. Were they not part of the web?

[BerislavLopac]

This all depends how we define the Web. If we focus on transport protocols like HTTP, then you are correct; but the whole point of the article is that you don't need browser/HTML/CSS/ajax/whatever to consume the content.

All that is no news, and the whole concept was very hip back in the days when XML and Web services were all the rage, some ten years ago. In my opinion, it's not about who will take over -- we will simply have different forms of content transfer over the Internet, and will decide on a case-by-case basis which is better in each particular situation.

[troystribling]

One big difference is that a lot of people are willing to pay $0.99 for and app while the same charge for a web page bookmark would not go over well.

[MisterWebz]

Excuse my ignorance, but i thought Ajax sends the whole page but only renders a part of the page?

[coderdude]

The server sends a client a whole page and once that page renders in a browser then you can use Ajax to transfer arbitrary bits of data to and from the server. You typically would not send an entire page to a client via Ajax, but you would use it to tell the server that 'User A' has voted on 'Story B' without having to load an entirely new page or refresh the page you're on.

[chc]

If your Ajax code is reloading the whole page every time anything changes, that's some awful code. I have seen such designs in the wild, but that's just because whoever wrote the script was a lousy developer. It's not supposed to be that way.

Here's how Ajax works: A whole page is downloaded once from the server, then the page's Ajax functions start running — and then when part of the page needs to be updated, the Ajax toolkit only has to reload that specific data and render it on the page.