Messages On Startup
Kevin Darcy
kcd at daimlerchrysler.com
Tue Aug 31 06:17:59 UTC 2004
phn at icke-reklam.ipsec.nu wrote:
>Kevin Darcy <kcd at daimlerchrysler.com> wrote:
>
>
>>Barry Margolin wrote:
>>
>>
>
>
>
>>>In article <cge2go$1ub7$1 at sf1.isc.org>,
>>>Kevin Darcy <kcd at daimlerchrysler.com> wrote:
>>>
>>>
>>>
>>>
>>>
>>>>Barry Margolin wrote:
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>>In article <cg6c75$nm3$1 at sf1.isc.org>,
>>>>>Kevin Darcy <kcd at daimlerchrysler.com> wrote:
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>>Hmmm, okay, I'll go tell our plant-floor folks that they can't use their
>>>>>>paint-control/milling/stamping/machining/welding/electronics-testing
>>>>>>devices any more and they'll just have to improvise somehow...
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>What's your problem? Just put "check-names master ignore" in the
>>>>>options section and you'll be all set.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>My only point is that a default setting of "fail" would be rather
>>>>Internet-biased and misguided.
>>>>
>>>>
>>>>
>>>>
>>>Another point in favor of that default is that it's a safer setting. If
>>>you're connecting to the Internet and don't have things configured in
>>>the standard way, you can cause problems for others. So it's best to
>>>have the defaults correct for the interoperation cases.
>>>
>>>If the default doesn't match your needs for private use, they only
>>>inconvenience you, not anyone else.
>>>
>>>
>>>
>>I can sort of see that point, Barry, but as I've already asserted in
>>this thread, it's usually large organizations that host DNS,
>>organizations that can be expected to have hardened processes that
>>prevent interoperability-causing data to be loaded into any nameserver
>>at all. So for that small category, a conservative check-names seems
>>rather superfluous. I would also point out that such large organizations
>>have an *incentive* to be as interoperable as possible, since more
>>interoperability means more visitors to the site(s), more interest in
>>the products, more sales, more revenue, etc. So if underscores cause
>>interoperability problems -- and I still remain rather skeptical about
>>that assertion -- then those orgs are going to crack down on
>>underscores, and if they have any brains at all, they'll stop the
>>underscores in a way that doesn't involve bringing down the whole zone
>>(which is basically the blunt-instrument approach that "check-names
>>fail" takes).
>>
>>
>
>
>
>>
>> - Kevin
>>
>>
>
>I don't see why underscores should be used AT ALL, there have been
>at various times problems, it _is_ against RFC. Why use something
>that _might_ impare when other characters are available ??
>
Aesthetically, I don't particularly like underscores either, but lots of
folks do, and don't give a rat's ass about purely-theoretical
interoperability issues. And as long as the (internal or external)
customer is paying the bills, how am I, or any DNS admin, in a position
to say "no"?
- Kevin
More information about the bind-users
mailing list