Is server-side user agent detection/sniffing bad?

Question

Client-side user agent detection is known to be bad and discouraged in favor of feature detection. However, is it also bad to react differently based on the incoming user ag

Answer 1

In "standard-browser" scenario it is not bad BUT it is not reliable since a lot of browsers offer the user some config option/plugin/whatever to modify the user-agent.

In such a situation I would implement something similar to facebook - they detect based on UA (and possibly other things a.k.a. as "fingerprint analysis") whether to redirect to a mobbile version (i.e. http://m.facebook.com/...) or not (i.e. http://www.facebook.com...). At the same time they offer a URL param m2w which overrides this redirection mechanism.

Depending on the mobile carrier it might even happen that they have some content-aware proxy/cache which does scale/recompress images on-the-fly and appears on your end as a "normal" borwser...

Thinking of scenarios outside the browser... for example if you are serving some specific protocol (like WebDAV) this might be the only option to have some sort of "platform-specific" behaviour (for example the difference between OS X and Windows).