[fprint] Outstanding issues

[Daniel Drake]

Eddie Hung wrote:
> I'm a recent convert to libfprint - after having tried pam_bioapi and
> pam_thinkfinger on my IBM Thinkpad, and I must say that, even at this
> early stage of libfprint, it seems to work a lot better than the other
> two solutions - and for that I must commend the author!
> 
> Allow me to make a few comments regarding things that aren't quite
> working as well as expected. I have done a quick search on the
> archives, and they don't seem to mention some of the issues I'm about
> to outline here.

pam_fprint is the component you are using with all of these 
applications, and pam_fprint itself is just a proof of concept:
http://www.reactivated.net/fprint/wiki/Pam_fprint

I'm not expecting things to work very well at the moment. I'm still 
working hard on the very lowest level of the system - fingerprint 
scanning library code and device abstraction - and have not yet focused 
on the bigger picture. pam_fprint was something I knocked up in 15 minutes.

It seems that the applications you mention (and perhaps PAM itself) are 
not really ready to move away from the "enter username then password" 
approach.

It is within my goals to produce a decent and comprehensive fingerprint 
authentication solution, but it likely won't look much like it does at 
the moment. Those tools will need to be reworked for a new interface.

Before I move onto implementing a more suitable authentication 
architecture and modifying desktops to use it, there is other library 
work to be done first which I am working on. When this is done I will be 
returning to the task of application integration - and doing it 
properly. It will take time, but it's coming :)

Thanks for the feedback,
Daniel