On 7/20/07, Les Mikesell <lesmikesell@gmail.com> wrote:
> Karl Fogel wrote:
> > Les Mikesell <lesmikesell@gmail.com> writes:
> >> Erik Huelsmann wrote:
> >>> I work in a financial institution. Users make mistakes. We correct
> >>> them, but the original error stays in the database, forever.
> >> What would you do if a user included something in your database that
> >> would be illegal for you to store or redistribute to anyone else? And
> >> I doubt if your comparison really comparable in other respects. Can a
> >> user accidentally include a DVD image in a database update instead of
> >> a small change?
> >
> > I think you might be misunderstanding Erik's point.
> >
> > He's not arguing that it's okay for errors to stay in the database
> > forever. He's saying that users make mistakes, and that even though
> > they can correct those mistakes for the future, the original mistake
> > stays around forever, and that's a *bad* thing. He's on your side :-).
> > (As am I.)
>
> I took it the other way. Financial transactions need a real audit trail
> so there's a good argument for saying that mistakes and their
> corrections should be tracked forever.
This is where I am as well. I work for a company traded on the NYSE,
which means we're subject to lots of Sarbanes-Oxley review. My
auditors would not be pleased if there was an easily-accessible way
for people to permanently remove things from Subversion - they were
absolutely giddy when they heard that that people (other than the
repository admin, who needs to go to quite a bit of trouble) can't
delete elements AND their history.
if I lost this measure of control, their faith in my change-control
procedures and monitoring would be shaken.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Sat Jul 21 18:23:06 2007