Section 11.5. Avoiding Content Rot

11.5. Avoiding Content Rot

Nothing endures but change.

Heraclitus of Ephesus (c.535-475 BC)

Creating simple web pages and making some of them dynamic and searchable is actually the easier part of setting up a good web site for a project. The harder part is making sure that the information stays correct. As soon as someone has written down the steps to do something, the process may change and make the web page incorrect. Such "content rot" may seem at times to be as inevitable as the second law of thermodynamics. The key ideas of this section are how to help project members remember to update the information, one piece at a time. Ideas to help keep the information on the web pages accurate are:

A designated librarian

Someone has to want to have accurate information on the web site. This person is acting as the librarian for the project. It could be a manager, a writer, or even the toolsmith. This person should monitor the content changes in the web site to make sure that links work and page formatting is consistent. If he is involved in the development and testing of the project, then he may know who to ask for updates to pages. When email is sent to the team about some change, the librarian can be the person to send back a request or reminder to update the related information on the web site.

Easy updates

The process to change a page has to be as easy as possible. This is the advantage of Wiki-based web pages, where the Edit button is right next to the content. Failing that, you can add text to the bottom of every page containing the name of the file for the current web page and the commands to check it out using the SCM tool.

Timestamps

Sometimes just knowing when a page was last changed is enough to decide whether its contents should be updated. The date can be found from your SCM tool, rather than relying on people to change a date by hand. Copyright dates are usually unhelpful for this purpose. Extracting the name of the person who made the last change from the SCM tool can provide a good starting contact for making future changes.

Templates

A small set of template files for web pages helps people add new information to the web site in a consistent manner. The templates can also contain comments about the steps to follow to add a new page to the web site.

Another approach is to add empty web pages for people, but only if they let you identify them on such pages in a large font stating who is responsible for adding content.

Link checking

A sure sign of a decaying web site is when all the interesting links don't work. You can use one of the open source link-checking tools, such as linklint (http://www.linklint.org) or checklint from W3C (http://validator.w3.org/checklink), to check that the URLs referenced on your web site aren't broken. These tools can also be run as part of your automation environment.

Web logs

Many web servers can tell you from their logs which files are most visited, from which you can create your own list of the files that are almost never visited. These files are often the ones that have become useless and need to be updated or discarded. Note that this doesn't refer to those cheesy counters that appear on some dusty home pages.