concerning devices (and networks, too) you are very right: It's pretty wild out there and it takes a lot of detail to make sense out of it.
Most of our device detection goes beyond the user-agent string to accurately detect devices. At the moment, we are able to identify 99,8% of devices, even if the request was proxied (via Opera, for example) or transcoded (via Novarra for example).
Comments
Nicely done. I like the name and the approach.
Is accuracy a problem? My understanding is that there are operators that configure devices to obscure the true user agent.
Hi eli,
concerning devices (and networks, too) you are very right: It's pretty wild out there and it takes a lot of detail to make sense out of it.
Most of our device detection goes beyond the user-agent string to accurately detect devices. At the moment, we are able to identify 99,8% of devices, even if the request was proxied (via Opera, for example) or transcoded (via Novarra for example).