Re: IANA type code registration clean up

[Edward Lewis <Ed.Lewis@neustar.biz>]

Title: Re: IANA type code registration clean up

(Answers to Bill's, Danny's, and Mark's replies.)

At 4:42 +0000 6/14/07, bmanning@karoshi.com wrote:

>On Thu, Jun 14, 2007 at 12:08:39AM -0400, Danny Mayer wrote:

(Bill)

>       So Ed is concerned that the IANA registry points to documents
>  and people who (might) change/move over time and the original
>  reference will be lost ... the IANA has a dangling pointer.
>

>       Is this the jist of your concern Ed?

Not exactly.  Although I have been labelled anal retentive [1] about keeping things in order, I agree that would be too much work.  The jist of mu concern is ... if I were to walk off the street and try to write up a DNS protocol element (server, stub, etc.) can I find all of the documentation to create a compliant hunk of code.

IANA's role, as registry of protocol parameters, is that it is the authority on things like the standardized RR types - much like ARIN will let you know who is responsible for an IP address or my employer would let you know who is responsible for a name registered with us.

IANA's registrations ought to be kept up to date as much as ARIN's.  With the case of ARIN, we expect the registrants to do this work, not the registry and I think it is fair to expect the same with IANA.  (That is why I am pressing this within the DNSEXT working group and not throwing flames at IANA.)

So, the job is on us to clean this up and hand the work to IANA.

>       Back in the day, I understood the IANA took the stance that
>    they would document assignments but it was not in their charter
>        to hold copies of the defining documents... that was the role of
>       the hosting organization, be it IETF, ITU, ATMF, BBN, MIT, etc.
>        If the hosting org chose to let the reference document lapse,

>       so be it.

I agree with that, IANA is a registry, not the Library of the Internet.  IANA is also not responsible to make sure that the entries in the registries they run are accurate, well-thought ought, etc.

My recent spate of email isn't about cleaning up IANA, it's making sure that a person can come in and find the resources to create an interoperable implementation without having to read BIND's source code or buy more beer for any one of us that do know the protocol.

Ok, so my main motivation is to lose weight, okay?  I admit it, I need to lose a few fat cells.  This really came from my medical care professional, still not from my tendency to be anal retentive.

(Danny)

>> Weren't other types like MAILA, MD, MF and WKS deprecated somewhere? I
>> can't find a document on it. If so then it's been done before and the
>> same path can be followed.

Yeah, I've heard the same rumor. ;)  I've also been asked for a reference document on "forward zones" and other such (Loch) Nessie[2]-like creatures of the DNS protocol.  (They are implementation specific configuration elements, not protocol standards.)  In recent years I have come to be surprised how little (percentage wise) of the DNS protocol is really in a document, instead residing in our brains.  'Course, it is good for job security.

But once we all retire, will the next generation of engineers be able to keep it up?

>       There may be a handful of people who -KNOW- why these RR codes were
>    created.  The documentation for their creation is, to my knowledge,
>    nonexistant.  Now the IANA has flagged the codes as [IANA-reserved],

>       but there is no data on how they got there, other than hearsay.

Ahh, Bill, perfect timing.  That's exactly the condition I want to reverse.  There's no reason not to document something and say here's how it is now.  (Kinda like the testament of the "Church of What's Happening Now.[3]")  This handful of people no doubt will want to go away and no longer be bothered with their youthful indiscretions.

Just because something came and left tracks doesn't mean we have to preserve it, we should document that these things are no longer in the mainstream.  It's alright to leave them reserved for some unknown, undocumented reason but let's put something down in writing that the class of 2173 engineers of La Univeridad del San Manning can use to reclaim the numbers when needed to allocate the VIXIE resource record.  By then, I'm sure there won't be one computer still around on the face of Mars still using the MB record.

At 16:29 +1000 6/14/07, Mark Andrews wrote:

>
>      I think the best solution for ATM would be for the IETF

>       Trust to get permission to hold and distribute the pdf.

Something to consider.  Who is the contact for that request?  Ray Palletier?

[1] http://en.wikipedia.org/wiki/Anal_retentive

[2] http://en.wikipedia.org/wiki/Nessie

[3] http://en.wikipedia.org/wiki/Flip_Wilson, "His characters included Reverend Leroy, pastor of the Church of What¹s Happening Now"

--

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Edward Lewis                                                +1-571-434-5468
NeuStar

Sarcasm doesn't scale.