Re: To the bind maintainer
On 23-Jan-00, 16:51 (CST), Greg Stark <gsstark@mit.edu> wrote:
> All machines should run a named.
Huh? Where does that come from?
> It's a new phenomenon that lots of machines on the net are windows
> or mac machines for which there just isn't a good name server
> implementation. Previously name servers were largely only for serving
> remote clients.
Its a very old (well, in internet terms...) phenomenon that a bunch of
Unix (and other OS) workstations, each quite capable of running named,
simply pointed there resolver at one or more machines that provided
answers.
What's the advantage of each machine having it's own named? Are you
saying that each should be a secondary? That becomes a pretty heavy load
on the primaries. Or are you arguing that each should be caching only?
Then each machine has to build up it's own cache...which means you have
a heavier load on everyone. (Do you also think every machine should run
it's own webproxy? Why not?)
> I'm a little annoyed at the logic that seems to be at work here. We're
> dummying down the average linux workstation with the argument that "if
> people don't need it then they're better off not having it".
That's weird. One of my objections to the MS model is that it keeps
putting stuff on my PC that I don't need or want. :-)
I object to the assumption that we need to adjust every package, even
those that are not aimed at newusers to cater to the ignorant. There are
appropriate places to spend the effort: net setup, X setup, etc. Putting
a bunch of Q&A in the postinst of named seems to me counter-productive.
Any one who can figure out how to set up a nameserver is probably going
to want mess with the config files any way.
The exception is the person who needs/wants local caching because of
being on a slow link; they'd probably be better served by a dedicated
dns cacher, but until we have one packaged up, having bind ask "do
you want a caching only server?" and doing the right thing might be
worthwhile.
> That's misguided and it's the kind of thinking that leads to Windows
> or Mac solutions where everything looks really pretty but a lot of
> functionality isn't there when you discover you do in fact need it.
The strange thing is that I think you (Greg S.) and I basically agree,
but we have very different opinions about how to avoid "dumming down".
Steve
--
Steve Greenland <vmole@swbell.net>
(Please do not CC me on mail sent to this list; I subscribe to and read
every list I post to.)
Reply to: