On 4/1/2015 11:05 AM, Stewart Bryant wrote:
> On 01/04/2015 18:38, Joe Touch wrote:
>>
>> On 4/1/2015 1:14 AM, Stewart Bryant wrote:
>>> Because the IPv6 delivery header does not include a checksum of its
>>> own, it is subject to corruption. 
>>>
>>> SB> It is subject to corruption whether or not it has a checksum.
>>> SB> The point is that there may be undetected corruption. However
>>> SB> detection in only probabilistic even with a checksum. So
>>> SB> I think that that this text should be:
>>>
>>> Because the IPv6 delivery header does not include a checksum of its
>>> own, it is subject to higher probability of undetected corruption. 
>> The probability of undetected corruption is 100%. And a checksum isn't
>> the only way to detect errors.
>>
>> IMO:
>>
>> Because the IPv6 delivery header does not include error detection, it is
>> subject to undetected corruption.
>>
>> Joe
>> .
>>
> I accept that there are other possible error checking technique, but
> the original text is talking about checksum. However no check is
> perfect so I still think it needs to be:
> 
> Because the IPv6 delivery header does not include error detection,
> it is subject to a higher probability of undetected corruption.

Point taken. Better.

Joe

_______________________________________________
Int-area mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/int-area

Reply via email to