Monday, August 30, 2010

The Web is dead?

That's the conclusion Mark Day claims is now made by Chris Anderson of Long Tail and Wired fame.

Day uses the model to construct a "media centric" view of the future of the Internet as being a home for "apps" - and the idea that newspaper proprietors are moving from free-on-line to "app" models of news distribution. He summarises it nicely as;

Conceiving, making and marketing apps is a creative entrepreneur's dream with low barriers to entry and global distribution available in an instant. But in the great majority of cases the best applications will be snapped up by the major media companies as they strive to deliver their content by covering all bases and meeting all consumer demands, all at once and all the time.

Now my earlier post on co-operation policy should explain why I actually think we should actively resist the idea that major media players will monopolise the app space. In part the history of the Web versus proprietary on-line services (think AOL) is part of the reason - there is more value created from ay-to-any connectivity than from walled gardens.

The actual original story in Wired is accompanied by a highly misleading graphic that shows proportions of internet traffic and shows the Web as a declining proportion, but so too e-mail. Yet in his argument Anderson still classes e-mail managing apps as part of the change (the dumb example being that an iPad can receive e-mail - I don't know why the iPad doing it is an app when my PC using Outlook isn't it). The proportions are misleading because almost by definition IPTV and P2P downloads are bandwidth consumptive. They are also services that will ultimately demand localised caching.

Anderson quotes Metcalfe's Law - that the "value" of a network is proportional to the square of the connections (or people connected). For a man famous for the long tail argument - which morphs into power laws - this is a dramatic error. The error is that it assumes that all connections are of equal value to any person. But the value of my connection to the other n-1 people in the network is likely to be distributed like a Power Law - the conclusion of which is that the value of a network (see 1 or 2) is proportional to n*log(n) where n is the number of connections. This doesn't mean that networks don't tip but they don't tip as fast.

It also demonstrates why building a flat network with the same amount of connectivity between every point would be wrong. (This concept was promoted by George Gilder in Telecosm which followed his Microcosm, an application of Havyatt's Law.)

Meanwhile elsewhere in the news Ericsson has been pitching the idea that mobile network operators need to come up with new ways to bill to allow users more affordable capacity. If Anderson were right the mobile operators wouldn't face this problem because their "content" and "apps" would all sit inside a manageable space. But they don't - basically what users access is still TCP/IP traffic - most of it delivered as HTTP.

So if the Internet isn't dead, and the Web isn't dead, can we agree with Anderson that Netscape cofounder Marc Andreessen was wrong forecasting the rise of the Web browser when he forecast the Web replaced PC application software and operating systems were merely “poorly debugged set of device drivers,”?

Well, no, because these Apps of which he speaks are all "outgrowths" of the net-centric world that was the browser. The Web is not dead, it has just become multi-faceted.

No comments: