Last modified: 2011-03-13 18:05:19 UTC
Depreciating older versions can result in loss of
transparency and subsequently mislead onlookers.
As this a somewhat complex scenario, I'll explain
using an example. Consider the following sequence
of events (which is btw the abstraction of a real
- revision 1 is reviewed by user A
- vandalized revision 2 is created and reviewed by user B
- a typo is corrected in revision 3, reviewed by user C
- user D removes vandalism in revision 4
- B depreciates revision 2
Obviously, user B is the one who did the sloppy work;
the revision history however doesn't reflect this:
revision 4 user D [sighted by User D] (remove vandalism)
revision 3 someone [sighted by User C] (typo correction)
revision 2 vandal
revision 1 someone [sighted by User A] (initial review)
Anyone not checking logs would conclude that user C
erroneously reviewed the diff rev1-rev3, when in fact
s/he did nothing but review the typo correction rev2-rev3.
As this will inevitably lead to people barking up the wrong
tree and since I fail to see any actual advantage with
depreciating old versions, I'd strongly suggest to just
deactivate it and allow depreciating the current stable
I don't like the idea of making it impossible to unreview old revisions. Seems icky. And what if vandalism snuck in, or an error was found in an "excellent" revision and someone wanted to lower its status?
I'd prefer some of the log tweaks over this.