In an ideal world sites would be evaluated based off of the relevant value of the content they contain. For many years now there has been a movement of folks who have a solution; the project is called open directory, and involves having actual trusted human beings making decisions about the value of a website. Some issues with the open directory project:
- There are billions of pages out there, tons being added every day
- There isn't a viable profit model for it, with the amount of human resources required. So it's handled by volunteers, folks who are working for nothing more than the benefit of others.
- The time it takes for a human to determine the relevance of a site is vastly greater than the time it takes a computer program to determine it, and a computer can check a lot more data points in that time. As it is now, we often wait 30+ days for Google, Yahoo, or MSN to spider our sites, mostly due to the massive amount of content, much of it changing on a daily basis.
Programmatically determining relevancy correctly, is not an easy task. Add in thousands of folks spending hours on end trying to think up ways to exploit the system for their own gain, and it gets even more daunting.
Backlinks are essentially references. References are a concept that has been around a long time, and used in many other areas of life to determine the trustworthiness of an individual. When determining trustworthiness of an individual, the amount of initial trust awarded is often based on the trustworthiness of the person vouching for the unverified person. Is it a perfect solution to match this system to determining relevancy for web sites? No. But, it does seem, at least to me, to be a reasonable sustainable source of programmatically verifiable data on which to make a decision regarding relevancy. It's only a single data point, perhaps it's weighted the most, but there are many other factors that contribute. It's tough for folks to exploit, easy for a computer to verify.
I agree with the spirit of your post, but I can't make an argument for a better viable solution. If someone could come up with a much more foolproof method of verifying relevancy, which could not be easily exploited by others, I can assure you that search engines would do everything in their power to implement it. There is very little they care more about than "true" relevancy, because increasing relevancy directly increases both profit and market share.
