This past week, I’ve been spending my time installing software of various sorts for my new job. Much of that has involved doing new things, including finding documentation for things that I need to do online.  In the course of this, I’ve run into a limitation of Google.
Much of the software that I’m installing is relatively new, meaning written within the last five years. And as popular open source projects, they’ve undergone frequent updates. And the documentation is not all that great.
But there’s been plenty of people who have shared their experiences in blog posts and others. And so you’d think that Google would be a great resource to find stuff.
But Google has a big problem in these cases. One of the major components of Google’s search algorithms is link traversal. A page that has many links to it is rated higher than one that has no links. But older pages are far more likely to have links to them – after all, they’ve been around longer for people to find and link to.
As a result, Google search results tend to skew to older pages. Which is a big problem if you’re looking for information about something that has been around for a little while, but has changed much in that time. The resources that you find are likely to be out of date.
You can go into advanced search and limit results by date. But I’d like to see all results, just skewed to the most up-to-date data.
Which goes to show, even the mighty Google is not the perfect solution for all problems.