There is another issue which no one seems to have mentioned in this thread: transparency
Assuming that this effort succeeds (!) it is inevitable that someone at some time will find some problem/bug/backdoor which will have been around for a very long time (see openSSL :-) )
At that point it will be vital that every halfwit "internet expert" be able to trace the origins of THAT particular line of code and the reasons for its change/introduction! Given that information anyone who can read code should be able to see if the change was a benign mistake or was in fact a a subtle "bad guy" attack.
Conspiracy theorists of course will still bather, but it is essential that the baselessness (SP?) of their blather be obvious lest they actually gain some credibility.
Mike On 06/17/2014 03:45 PM, Alain Forget wrote:
And here's an example difference between blame with TC history: https://github.com/discnl/truecrypt-source-archive/blame/master/Common/Crypto.c and without: https://github.com/CipherShed/CipherShed/blame/master/Common/Crypto.cHm, excellent example of how keeping in the history appears more messy. However, in the left-most column , it seems to clearly state which version (of TrueCrypt, I assume) that was last changed/added, so we can ignore changes before a particular version (say 7.1a, most likely). However, this still isn't terribly usable. I'm familiar with subversion, where we can compare specific revisions/commits to each other. Is it possible to do something similar in blame, whereby we start from only a particular version and see all the changes up to the most recent (or within some range), so blame since 7.1a (or whatever version) up to the current version? If so, that would mostly solve this issue, although it might still be a slight pain to always have to remember to blame starting from 7.1a. Alain
-- *"Ayn Rand - Philosopher in Chief to the intellectually bankrupt" Sherlock Holmes* This work is licensed under a Creative Commons Attribution 3.0 Unported License.