Re: [Last-Call] Artart Last Call review of draft-ietf-tsvwg-ecn-encap-guidelines-20

[Bob Briscoe <[e-mail]>]

Paul,

Thanks for the reassurance. See [BB]

On 07/11/2023 15:49, Paul Kyzivat wrote:
> On 11/7/23 7:42 AM, Bob Briscoe wrote:
>> Paul,
>>
>> Thank you for checking this. See [BB]
>>
>> On 25/10/2023 19:54, Paul Kyzivat wrote:
>>> Document: draft-ietf-tsvwg-ecn-encap-guidelines-20
>>> Reviewer: Paul Kyzivat
>>> Review Date: 2023-10-25
>>> IETF LC End Date: 2023-11-02
>>> IESG Telechat date: ?
>>>
>>> Summary: This draft is basically ready for publication, but has nits 
>>> that should be fixed before publication.
>>>
>>> This is a clear and well written document. I have no substantive 
>>> issues.
>>>
>>> I ran the IdNits tool and it reported a number of things that should 
>>> be checked and possibly corrected. I won't repeat most of them them 
>>> here.
>>>
>>> Among them were a number of lines with non-ascii characters. I 
>>> investigated these and all seem to be unintentional - apparently a 
>>> non-printing character.
>>
>> I can never understand whether I am meant to take note of these 
>> non-ASCII warnings.
>>
>> Like a number of the drafts I've been taking through to RFC recently, 
>> this is a v old draft, that was written in xml2rfc v2. My wysiwig XML 
>> editor stlil only supports v2. So at the end of WGLC, I converted the 
>> doc to xml2rfc v3 (using the author tools) which replaces all the v2 
>> XML entities like   with the non-ascii equivalent. So then, I 
>> manually tweaked the v3 XML necessary to define the various XML 
>> entities for the non-ascii characters again, and replaced all their 
>> occurrences. Then, when the author tool produces the txt, it outputs 
>> all the instances of XML entities again to non-ASCII characters in 
>> the txt.
>>
>> What is wrong with that? And I thought I was following the guidelines 
>> (I also asked Jay and followed his advice as to which online advice 
>> to follow)? So am I meant to follow the guidelines, or am I meant to 
>> take note of the warnings in the nits? I'm perplexed.
>
> There may be *nothing* wrong here. IdNits is calling it out, but 
> sometimes it is over-zealous.
>
> OTOH, in this case I *think* the non-ascii chars are some sort of 
> whitespace. If these are indeed &nbsp and that is important then I 
> have no issue.

nbsp was just an example. This draft uses others that are not 
whitespace; specifically, all of:
<!DOCTYPE rfc [
   <!ENTITY nbsp    "&#160;">
   <!ENTITY ouml     "&#246;">
   <!ENTITY ndash     "&#8211;">
   <!ENTITY mdash     "&#8212;">
]>

I don't understand why we have tools that correctly produce non-ascii 
txt, and author instructions on how to use them, but then other tools 
warn us not to do this. Is this just that idnits hasn't caught up?


Bob

>
>     Thanks,
>     Paul
>

-- 
________________________________________________________________
Bob Briscoe                               http://bobbriscoe.net/