Re: [AVTCORE] Errata 4192 RFC 3550

[Julius Friedman <[e-mail]>]

And I see that 3551 obsolete 3550.

This is also a trivial change just as are the suggestions which are
outlined in the latest changes which have not yet made it into the rfc e.g.
the -new documents which bears the same rfc number but changes wording.

Anyway we have other problems to worry about  besides this.

All I know is that my application will not comply with this old way of
thinking and without having the source code one person won't even really be
able to tell why the calculation is giving anyway.

I have already shown in my own server implementation that existing software
works the same if not better using my values as recommended.

I will write a new draft regardless of what you do with this errata but I
suggest it is not rejected without proper reason.

...

Please also check out the errata related to rfc2326 and lets try to get
somewhere with that...

Sincerely,
Julius Friedman.
On Dec 11, 2014 10:07 PM, "Tom Taylor" <[e-mail]> wrote:

> RFCs are not updated (with the exception of errata, which are more in the
> nature of trivial corrections); they are replaced. As a present example,
> the original RTP specification was RFC 1889, published 18 years ago. If you
> go to the RFC Editor database,
>
>
> http://www.rfc-editor.org/search/rfc_search_detail.php?
> rfc=3550&pubstatus[]=Any&pub_date_type=any
>
> you'll see that RFC 3550 obsoletes RFC 1889.
>
> Tom Taylor
>
> On 11/12/2014 4:33 PM, Julius Friedman wrote:
>
>> Thanks for taking time to... inform me...
>>
>> And so then what are the modifications to the standard proposed
>> considered? Is a modification not performed because of errata?
>>
>> This is not about xr reports or anything else besides the simple
>> calculations which obviously are not consistent.
>>
>> This is basic math and indeed proves if nothing else that the sender
>> octet count becomes USELESS when csrc, extension or padding is used.
>>
>
> [PTT] Useless for your purpose. I am telling you that your purpose is
> different from that of the people who accepted the calculations in RFCs
> 1889 and 3550 for the last 18 years. As a result, a proposal to replace RFC
> 3550 for this reason won't, honestly, get anywhere, but if you really need
> that information you can probably get it using the RTCP XR framework.
> ...
>