Re: netwrk stuff

[Douglas Otis <[e-mail]>]

On Jul 24, 2006, at 8:06 AM, Todd Glassey wrote:
> On Jul 24, 2006, at 7:24 AM Douglas Otis wrote:
>>
>> The completion of documents, and the closing of WGs remains within  
>> the competence of the IETF.  Beyond describing the intended use  
>> and the vetting initially achieved, there is little benefit  
>> derived revisiting a document unless a change is required.
>
> How so ? - this is a statement of objectivity or the lack of it  
> more accurately.

This was attempting to describe what remains within the realm of IETF  
competency.  Attempting to define which iteration of a protocol  
(based upon the introductions of RFCs) is needed for interchange  
within some distribution, or a company's application, or some OS  
would be objective and likely lacking accuracy when done by the  
IETF.  After the software becomes fully debugged, the Current  
protocol iteration might become Stable, but prior to compliance  
testing, each company or distribution should declare which iteration  
is providing Stable interchange.  This declaration is beyond the  
competency of the IETF, and should the IETF attempt to make  
declarations of this nature, such would likely be based upon the  
information provided by dominate companies or distributions.  Here  
the IETF becomes the IVTF.  RFCs represent a continuum of change.   
What is declared as Stable by distribution X will likely be of far  
more value than a similar assertion attempted by the IETF.


>> When a change is required, the affected document should be marked  
>> as updated or replaced, where again the intended use and the  
>> vetting initially achieved should once again be noted.
>
> There may be issues with that. One of the things that the IETF has  
> not dealt with is the use of the IETF's processes as weapons  
> against others with less presence or power within the IETF WG's -  
> the problem is that not everyone here is a nice-person - some are  
> actually really bad people who are here intentionally to screw  
> others who don't agree with them or who have opinions that are  
> dangerous to initiatives or the the political framework of the IETF.

Not every company fairly assesses a protocol either.  There could be  
a perceived advantage thwarting innovation when there is no intent to  
commercially make use of it.  Rather than thwarting innovation, these  
efforts could be diverted onto a separate track by utilizing a  
separate tracking overlay.  This would allow the market to decide  
whether there is value in adopting the newer track.


> Bluntly the IETF has grossly failed in being either Open or Fair  
> and Frank these Are what need to be addressed. The IETF's processes  
> are best described as a twisted-fantasy through loophole land. They  
> don't serve to make standards honest and fair but rather impossible  
> for a mere mortal to participate with and that is more offensive to  
> the world than the nasty language I used in describing the failures  
> of this and the previous IETF's in coming to grips with what the  
> IETF has become - a haven for career standards participants who  
> need to be somewhere to justify their status within their sponsors.

The SRD overlay approach permits separate development tracks.  This  
multi-tracking ability might avoid some innovations being seen as a  
threat.  The specifications for email, instead of being a free list  
of RFCs, becomes a compiled set of RFCs defined as Email.3.  Imagine  
that Email.3 is commonly seen as Stable, and Email.4 is Current.   
When a small enclave wants to introduce an incompatible change based  
upon Email.3, the SRD overlay could adopt a different reference name  
such as Foo-mail.1.   The present, somewhat simplistic, view that an  
RFC is updated or replaced assumes these RFCs support just a single  
track of change.

Even assuming a single track, the present system makes declaring what  
is being used difficult.  Knowing what is involved for a particular  
protocol is only clearly defined for the present set of changes.   
Standard categories are too poorly maintained to be useful.  A fair  
amount of effort is required to determine the sequence and dates of  
the various RFC documents to determine a prior set of RFCs.  When  
development involves collaboration between other organizations, being  
able to predict a reference earlier within the process would also be  
a benefit provided by an overlay approach.

The IETF should focus upon providing serialized documents that both  
support existing protocols as well as the introduction of newer  
protocols.  There is a need for the IETF to improve upon the tracking  
of their effort.  This tracking is also likely the only change that  
will eventually be found beneficial.  To declare something a  
"standard" assumes certifications and compliance mandates.  In many  
cases, until some experience is obtained, the IETF is unable to state  
that something can be implemented from the document and that it will  
be Stable.  Should the IETF even attempt to make and revise these  
assessments?  It seems that many of these assessments are dependent  
upon the beholder.  Provide a simple serialized tracking scheme and  
then pay attention to what your favorite OS or distribution declares  
as Stable, Deprecated, or Obsolete.

The IETF can avoid some potential trouble by retaining the  
Experimental categorization within the overlay.  Until enough vetting  
ensures the protocol will not damage the Internet, retention of this  
categorization indicates how this document is to be used.  The SRD  
draft provides Core, Extension, Guidance, Companion, and Experimental  
categories to indicate the intended use of an RFC.

-Doug






_______________________________________________
Ietf mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/ietf