Multi-language support

Many proxy vendors that support URL filtering also have the ability to rate the webpages that aren't in URL category database in real-time. They do this using automated programs that scan the web pages and categorize them based on the content of the website. Part of this is being able to recognize the words, tags, and tokens used on the website. That means the real-time categorization engine has to understand different languages if you have sites that world-wide. Foreign language support is probably not as wide-spread as you'd expect it to be with proxy vendors, so you should ask your vendor what languages they support in their real-time engine.

Some even have languages that aren't necessarily commonplace. For example, Blue Coat reports support for Klingon, as well as a made up language they call "Pornovian", basically common terms found on porn sites, for them to more easily rate a site as a pornography site.

Some vendors also report language support in both real-time engines as well as background rating engines (those engines used when the page doesn't have any obvious markers that would allow it to be rated in real-time). In most cases, the background rating engine will have more language support than the real-time engine.

So if your organization is truly multi-national or global, make sure your proxy vendor is as well.