[fprint] why enum fp_finger?

[Simon Kenyon]

Joshua N Pritikin wrote:
> I must be missing something really basic. Why is enum fp_finger part of 
> the API? Do people actually enroll all their fingers? If they do, why? 
>
> It seems to me that I want to be identified by a single finger for most 
> applications. Perhaps for high security, it would be worth testing more 
> than one finger?
>
> If identification frequently has false negatives then it seems like I 
> would want multiple versions of a single finger. Why does the API 
> directly support multiple fingers but offers no specific support for 
> multiple versions of the same finger? Or am I missing something?
>
> Wouldn't it make more sense to remove enum fp_finger and move it to the 
> application side (if needed)?
>
>   
bacause single finger is usually not good enough

# Fingers
	FAR
1 	~1%
2 	~1%
4 	~0.1%
8 	~0.01%
10 	<0.01%

Representative Finger 1:1 FAR Rates for FRR = 0.01% to 0.001%

The above table shows that with single finger, you will be 1 false 
accepts in 100
If you are happy with that, then go for it.
--
simon