Jean-Michel Pouré - GOOZE wrote: > > It is _expected_ to have a _highly_ variable time for prime number > > generators. > > This is understood now. So please let us remove key generation time on > the wiki for all cards.
This suggestion could easily be misinterpreted as promoting security by obscurity. Of course we want to have as many and as high quality details as possible about as many products as possible. > This is kind of spam to me to let users believe > that some card can do fast and the Feitian PKI is slow. A clueless user maybe. But then we should help give them a clue, rather than hiding any and all information that we think could confuse them. > I am quite surprised that people modified the Feitian PKI page and > added 2m38s and other cards have very low time. This is quite likely because of poor quality statistical analysis. You mentioned that you did 10 tests, which is already much better than just one test, but I agree with Ludovic that we should gather even more samples. Min max and avg times would be useful at the very least. > Some users might believe: Hey this other card is fast. My timings > were in the range of 1 minute. If a user sees those two numbers and makes a purchasing decision based solely on them, then they are really way out in orbit, and would need some serious education. It's not so much the task of OpenSC to do that education, but I fully understand that it is the interest of vendors like yourself to guide customers to the product that fits them best. However the OpenSC project isn't the very best conduit for that.. There's enough to do with the software here.. :) > This is all what I am asking for: equal treatment. This sounds a little like the SAS ground staff member who complained how it was not fair to others if I would be allowed by the cabin crew to take my carry-on into the small airplane, when I politely requested that they make the decision if my bag was too big. (It was not, which I knew already.) We are not in kindergarten. Life is not fair. Those who present convincing arguments will convince. On the other hand of course OpenSC should not have misleading or even worse false information! The best move would IMO be to do some thorough scientific testing of Feitian and put the numbers up on the wiki. Then that product will be the only one with actual great documentation. ;) //Peter _______________________________________________ opensc-devel mailing list opensc-devel@lists.opensc-project.org http://www.opensc-project.org/mailman/listinfo/opensc-devel
