[rfc-i] References

paul.hoffman at vpnc.org (Paul Hoffman) Tue, 23 February 2016 23:30 UTC

From: "paul.hoffman at vpnc.org"
Date: Tue, 23 Feb 2016 15:30:44 -0800
Subject: [rfc-i] References
In-Reply-To: <6FA9944B-DBB6-45E5-810B-0A3417B3ACB9@cisco.com>
References: <6FA9944B-DBB6-45E5-810B-0A3417B3ACB9@cisco.com>
Message-ID: <59D97957-0B24-478A-8EA5-02AF6B023EB0@vpnc.org>

On 23 Feb 2016, at 14:52, Joe Hildebrand (jhildebr) wrote:

> I'm double-checking the reference format in HTML.  Here are a couple 
> examples for author names:
>
> <author fullname="??"/>
> <author fullname="??" asciiFullname="English Name" 
> asciiInitials="E." asciiSurname="Name"/>
> <author initials="?." asciiInitials="N." surname="One" fullname="?o 
> One" asciiFullname="No One"/>
>
> All of these legal, and probable.  The v3 format does not require any 
> attributes in the author tag.
>
> I believe that draft-iab-rfc-nonascii-00, section 3.7, implies that if 
> there are *any* ASCII attributes, that a two copies of the reference 
> are inserted, the first using as many of the ascii attributes as 
> possible, the second using as few of the ascii attributes are 
> possible.  Is this correct?
>
> If so, we're going to need an algorithm for which attributes to use, 
> for both the initial authors (who in the current format is listed as 
> "Surname, Initials"), as well as for the final author (who in the 
> current format is listed as "Initials Surname")
>
> In my perfect world, we would *always* use the fullname, because 
> that's the only one that works with everyone in the worlds' names.  I 
> know that goes against the current style guide (and that style guide's 
> precedents), but none of those precedents really had to deal with 
> non-Western names.
>
> Even if we want to ignore all of the non-ASCII issues, Cher's name is 
> valid in the *v2* syntax:
>
> <author fullname="Cher"/>
>
>
> But I don't see how to render it using the rules in the style guide.
>
> Regardless of what we do here, we need rules that apply to all of the 
> output formats, in order to make sure all of the output formats look 
> as similar as possible to one another.


Fully agree with all the above, and note that the problem is not just in 
rendering references, but also author names on the front page. Further, 
there are questions about how to render things other than author names. 
Assume that the author of a draft is:

<author initials="P." surname="Hoffman" fullname="Paul Hoffman">
<organization ascii="Example Corp.">?x?mple Corp.</organization>
</author>

How would that be rendered on the front page of the draft?

Somewhat related: some authors might add "ascii" attributes reflexively 
even where they are not needed. Would anyone object if we added a step 
to draft-iab-rfcv3-preptool that checks if the ASCII attribute and the 
content are identical and, if so, remove the ASCII attribute?

--Paul Hoffman