Hi, i wrote: > > The reason why this is still not fully reflected by the man page is > > not yet uncovered.
Gene Heskett wrote: > Maybe a wee bit of security by obscurity? There is that I think in > everyones thinking on this subject. They don't want to price the farm > so cheap that it will actually sell. Ah no. The obscurity principle is unpopular in cryptography. The widely accepted method is to have the algorithms public, so they can be analysed and discussed, and to have the secrets separated in keys. Given that Theodore T'so can probably cause a text change in the man page if he really demands it, i rather expect to find a nitpicker like me who challenges the flat deprecation of /dev/random by some thin but valid argument. Just a gut feeling of mine. For my own decision of /dev/random against /dev/urandom: I use either of them very rarely. I have to deal with several old kernels of which i do not know how firm the opinions were when those kernels were young. So i will continue to use the legacy interface as long as it is available. But i will not raise objections if some day it becomes exactly the same as the /dev/urandom interface. This is the decision of the maintainers (Theodore T'so and Neil Horman of CRYPTOGRAPHIC RANDOM NUMBER GENERATOR), whom i deem more educated on the topic than i am. Have a nice day :) Thomas
