SELinux last straw
Les Mikesell
lesmikesell at gmail.com
Wed Oct 17 21:57:48 UTC 2007
Mikkel L. Ellertson wrote:
> Granted, the tools for SELinux are not as mature as the firewall
> tools, but does that mean we throw out SELinux instead of improving
> the tools?
No one is arguing that it should necessarily be thrown out. But, should
people be using it without understanding it?
> I have seen the same kind of arguments about just about every major
> change. I remember people complaining about udev, and what was wrong
> with using the standard /dev setup. I heard it about the change to
> IPTables. I have heard it about HAL. Way too many of them boil down
> to I know how the old system works, so why should I learn about this
> new way of doing things.
It's not just a matter of learning new things, and even if it were, that
would boil down to large sums of money in any business context. Think
about upgrading a large farm of servers that have multiple network
connections and the upgrade OS version detects the eth? devices in a
different order (real example, by the way...). Now you need the staff
at each location to either relocate the cables to match or edit a vast
number if ifcfg-eth? files after they somehow figure out what's
connected where.
> I am happy with the way things are working
> now. Don't change things and make me learn a new method. I don't
> care if this new method has advantages over the one I know.
Try it this way: there's been 30 years of work aggregating and improving
with the old assumptions. That's why we like unix-like systems. Do you
want to throw that out on the chance that an untested new idea might be
better?
> Now, some of the new things are not going to work out, or in trying
> to implement them, a better way may present itself. But if nobody is
> willing to try the new methods, and work out the bugs that are
> always going to crop up when trying something new, then there will
> not be any progress.
Research is always a good idea but most people want the testing to be
done before the new thing goes into production.
--
Les Mikesell
lesmikesell at gmail.com
More information about the fedora-list
mailing list