(Javier, I am CC:ing you to ping you for the last time, hoping you are
still alive and could provide some input on this issue.)
After seeing this bug stalling for monthes, I have looked at www.d.o's
CVS and tracked these statements back to the initial revision of
webwml/english/security/index.wml, committed in 1998 by James A. Treacy,
who looks MIA since August 2004. I am CC:ing him, but my guess is that
if we're lucky enough to get a reply, it will be along the lines of
"This statement was accurate in a previous era, but needs new
statistical grounding if the current maintainers still believe it is
valid.".
The data in Javier's study starts in 1999, where the mean time indeed
looks under 2 days. This quickly grows to 10 days in 1999.5, to never go
back down significantly until the end of the data collection in 2003,
where the mean time is 80 days. I do not think that, with a security
team growing proportionally to the archive, keeping the mean time under
2 days would be impossible, but unfortunately, the still recent
involvement of Moritz Muehlenhoff in the stable security team is the
only thing keeping it from looking like a one-man team. Debian should be
honest about this neglected side. Removing 2 sentences isn't really
hard. Not saying anything about the performance *is* better than saying
something wrong.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
