[ntpwg] Pending NTP WG Last Call on Autokey

[Danny Mayer]

Dave,

You can ignore the legal case referenced. I read the decision in its 
entirety (all 101 pages) and while it does reference timestamps and 
dates and times it does not have any specific requirements for time. 
Indeed for most purposes outlined in the decision the timestamp on a 
document residing on a computer could easily be 5 minutes off and still 
be acceptable, hardly an NTP issue.

The need for accuracy in the timestamp depends almost entirely on the 
nature of what the attorney is trying to prove. You and I both know that 
there are 100 different ways from Sunday to provide false timestamps on 
any document on the system without detection. Witness this odd quote in 
the decision: "it is common for the proponent to provide evidence of the 
input procedures and their accuracy, and evidence that the computer was 
regularly tested for programming errors." Noone is testing their 
computers in this way for the applications that they are running except 
for the people developing the applications. When was the last time that 
you tested Microsoft Word?

You can read the actual decision if you can stomach the dry legalese 
here: http://indianalawblog.com/documents/Lorraine_v_Markel.pdf
The discussion is mostly about the Federal Rules of Evidence and 
admissibility.

Now can we get back to the working group issues?

Danny

David L. Mills wrote:
> Todd,
> 
> I'd really like to understand your points. What are "first-person v 
> third-person type testimony models"? The court expressed the need for 
> improved resolution. The resolution of the PTP and NTP representation is 
> today better than the time light travels one foot. There must be 
> something else here.
> 
> The more interesting point is "absolute reliability in the veracity of 
> the logging therin". We computer science and engineering types interpret 
> that as impossible to do. We can provide massive redundancy and 
> diversity and clever mitigation algorithms at sometimes considerable 
> cost. The computer science community has tossed this issue around for 
> many years. We cannot provide "absolute reliability". Surely, competent 
> counsel and consultants should have made that point. Go talk to the 
> patent lawyers; they are engineers.
> 
> Now, where is the lesson here for the NTP community? You have predicted 
> dire consequences and meltdown. Be more specific. What is broken and 
> what is to fixed? We need an engineering colloquy, not a legal brief.
>  
> Dave
> 
> Dave
> 
> TS Glassey wrote:
> 
>> Rob - the Court Ruling from the US District Court in Lorraine V Markel 
>> (by
>> Judge Paul W. Grimm on May 4th 2007) set the standard for what digital
>> content and first-person v third-person type testimony models. This is 
>> now
>> something that the world is going to address - and it changes what the 
>> world
>> needs from timekeeping systems IMHO. What they need now is good or better
>> resolution - and absolute reliability in the veracity of the logging
>> therein. Without that - timesetting's are based on the credibility of the
>> person making that setting IMHO.
>>
>> Todd Glassey
>>
>> ----- Original Message -----
>> From: "Rob Seaman" <seaman at noao.edu>
>> To: "NTP Working Group" <ntpwg at lists.ntp.org>
>> Sent: Sunday, May 11, 2008 8:12 AM
>> Subject: Re: [ntpwg] Pending NTP WG Last Call on Autokey
>>
>>
>>> I'm concerned to see this bout of sniping - however, sniping was
>>> certainly one of the internet's earliest protocols. (One might,
>>> however, tender the observation that some people have expertise in
>>> both physics and computer science - and in systems engineering and
>>> human factors, for that matter.)
>>>
>>> Stripping away the ad hominem layer (as adding no signal to the
>>> discussion):
>>>
>>> TS Glassey wrote:
>>>
>>>> AutoKEY is not the only auth mechanism that NTP should support
>>> Returning to the beginning of this thread:
>>>
>>> David L. Mills wrote:
>>>
>>>>>>>>>>>>> There really is no wiggle room other than deprecating
>>>>>>>>>>>>> Autokey in its present form and reformatting the headers. I
>>>>>>>>>>>>> am not opposed to that in principle, but others,
>>>>>>>>>>>>> specificlly USNO, have not been heard from.
>>> Where is the disagreement? There is a world of difference in any
>>> standards discussion, of course, between what should happen and what
>>> will happen. The answer always lies in what needs to happen.
>>>
>>> One is confident that a single workable consensus - not compromise -
>>> remains possible.
>>>
>>>> broad-distribution models with this NTP is not possible.
>>>
>>> Clarification, please, on what is meant by a "broad-distribution
>>> model"? How, in detail, does broad-distribution differ from other
>>> kinds of distribution? That is - describe the problem before
>>> suggesting a solution.
>>>
>>> Does this thing called broad-distribution fall within the purview of
>>> NTP in any event? And is the world really clamoring for it? Does our
>>> world's social and technical infrastructure *need* "broad-
>>> distribution", or does this simply reside in the Hilbert space of what
>>> *should* be?
>>>
>>> Rob Seaman
>>> National Optical Astronomy Observatory