All the comments so far have been with regards to modifying a site's appearance based on the device.
A totally separate use case is for business intelligence reporting. It's not so bad with the iPad 2 vs iPad Mini because they have virtually identical hardware other than the physical screen size (same CPU, memory, CPU, etc). However it's more important with, say, the iPad 3 vs iPad 4. Although both are fairly similar (both are branded as the "new iPad"), they do have different CPUs.
For our business, it is valuable to know which physical devices people are using, not just which OS or browser version they are on. Our product is one which heavily relies on the CPU performance - knowing which devices our customers are on tells us which devices we should prioritise the testing on. Yes, we may be able to optimise the product to work on an iPhone 3G (picked as an example of a lower end iPhone), but if the number of customers who use that device is low enough, then the business case won't stack up. When it comes to device testing, it seems only prudent to try and mimic the device profile of our customer base.
Also there are other, perhaps less tangible, use cases for understanding the device profile for our customer base. For instance, if iPad Minis are more popular for a particular demographic (perhaps rush-hour commuters), this may inform business decisions about development priorities, feature roll-out, or marketing campaigns.
In summary - please don't assume that the only reason for device detection is to change the appearance of a site; the data can be valuable in other ways too.
A totally separate use case is for business intelligence reporting. It's not so bad with the iPad 2 vs iPad Mini because they have virtually identical hardware other than the physical screen size (same CPU, memory, CPU, etc). However it's more important with, say, the iPad 3 vs iPad 4. Although both are fairly similar (both are branded as the "new iPad"), they do have different CPUs.
For our business, it is valuable to know which physical devices people are using, not just which OS or browser version they are on. Our product is one which heavily relies on the CPU performance - knowing which devices our customers are on tells us which devices we should prioritise the testing on. Yes, we may be able to optimise the product to work on an iPhone 3G (picked as an example of a lower end iPhone), but if the number of customers who use that device is low enough, then the business case won't stack up. When it comes to device testing, it seems only prudent to try and mimic the device profile of our customer base.
Also there are other, perhaps less tangible, use cases for understanding the device profile for our customer base. For instance, if iPad Minis are more popular for a particular demographic (perhaps rush-hour commuters), this may inform business decisions about development priorities, feature roll-out, or marketing campaigns.
In summary - please don't assume that the only reason for device detection is to change the appearance of a site; the data can be valuable in other ways too.