Log in

No account? Create an account
On middlemen - Arvind Narayanan's journal [entries|archive|friends|userinfo]

On middlemen [Jan. 18th, 2010|08:28 pm]
Arvind Narayanan
[Tags|, , , , , , , ]

Here are a couple of comments I posted elsewhere, and I'm reposting them here because I noticed there's a common theme, which I will highlight.
Up until a few months ago I had the same reaction as most tech people on the subject of URL shorteners: they are messing with the Web and I wish they would die. Now I have a very different opinion. But first let me deal with the two common objections:

1. A URL shortener is an added point of failure. I wonder if people had a similar complaint back in the day when DNS started to become popular. When you add one more level of indirection you always add one more point of failure. I think it’s pretty likely that as the technology matures, the probability of failure will essentially vanish. We’ll see.

Linkrot in general used to be a big problem in the early days of the web, but doesn’t happen nearly as much anymore.

2. Short URLs hide information from users. If you’re building a consumer-centric website, the additional effort in expanding all the short urls is tiny, so I don’t see what the problem is here.

Now I get to the real point of this post: the one huge advantage of URL shorteners is that due to the extra layer of redirection, interesting things can happen at that layer, like publicly available click stats. Bit.ly already does this. Instant, public stats are going to be an incredible enabler of the real-time web. The significance seems to have gone more or less unnoticed so far, but I think it will become apparent in the months to come.
I've since noticed that goo.gl also plans public stats; whether or not the average web developer is aware of this trend, the big G is right on top of it.

On the topic of the Australian government forcing Google to censor search results there:
This issue has been popping up in different countries all over the world in the last few months. Most notably in China, of course, but also in India, France and Italy, where Google executives are in fact facing the possibility of jail time. See here and here.

The role of Internet middlemen in enforcing copyright was a key legal issue in the last decade. It appears that censorship will be the analogous issue for the next decade. We seem to have reached a relatively happy middle ground with copyright—middlemen have some responsibility, but a strong form of copyright protection has proved unenforceable, forcing many industries to innovate or die. We can hope that a similar thing will happen with censorship, with oppressive governments either collapsing or being forced to allow free speech.
I think I can distill several general ideas from these comments, most of which I hope will be controversial :-)
  • While the Internet has destroyed some middlemen (e.g., real estate agents), it has in fact created more middlemen.
  • One reason there are more middlemen is that these middlemen can be stacked on on top of the other, as in the bit.ly example.
  • Although Internet middlemen have haters both in the tech world and in the industries they replaced, I think middlemen are a good thing in general. They are good because they skim relatively little off the top, and because they can interface with each other in an automated way, without humans feeling the pain of dealing with multiple levels of middlemen.
  • Needless to say, they also tend to be vastly more efficient and cheaper than the pre-Internet middlemen they replace. In some cases, they might even play a role in fundamentally reshaping society and making it freer.
I certainly appreciate the dangers of middlemen, especially that they might become gatekeepers or monopolies. Indeed, I've attacked Craigslist on these grounds, and I don't like Apple's app store either. But the dangers need to be considered in balance against the benefit they provide, and opposing digital middlemen a priori because of the possibility of future abuse of power is silly.