Re: [apps-discuss] I-D Action: draft-ietf-appsawg-about-uri-scheme-00.txt

[Mykyta Yevstifeyev <[e-mail]>]

Hi Frank,

My response as document's editor.

18.10.2011 9:52, Frank Ellermann wrote:
> On 18 October 2011 08:09, Jiankang YAO wrote:
>
>>   Pls kindly help to review it before 25 Oct.
> Section 2.1:
> Please replace 'segment' by '*pchar' and import pchar
> instead of segment from STD 66.  The * indicates the
> main point "zero or more" (unlike<segment-nz>), and
> pchar is for me clearer than segment:  Above all it's
> none of the five terms for slightly different paths.

Well, RFC 3986 defines:

>     segment       = *pchar

So "/segment/" is no different from "/*pchar/".  I actually see no 
benefit from changing this definition, but if there is WG consensus on 
such change, it will be incorporated in the doc.

>
> The SPU gibberish is just ugly, please replace it by
> "registered token" or similar.

SPT is used not to say "special-purpose 'about' URI token"; and this 
implies the main purpose of creating *special-purpose* URIs/tokens.  I 
actually think this is the best though a bit obscure terminology.  
Please propose something particular in lieu of SPU/SPT terminology, if 
you want changes, and such proposals will be considered.

>    Ditto SPT.  Get rid
> of almost all MUSTard in this perfectly harmless URI
> scheme, only about:blank deserves a MUST.

Please provide particular cases of using MUST an justification for 
removing such MUSTs.

>
> For more than six months I try to get about:about in
> this draft, nothing happened.

about:about isn't appropriate for special-purpose URI:

>     defining an SPT should
>     only be used as a last resort approach, when a strict limitation on
>     handling 'about' URI with such SPT is necessary.

I see no such necessity for strict limitation.  One application might 
want to resolve about:about to wen page saying "I have no idea what is 
about:about", other will resolve to list of supported 'about' URIs, the 
third will do else, and I personally don't think we should unify 
handling of particularly this URI.  But, again, if there is WG 
consensus, I'll add the text.

>   Instead there are tons
> of obscure terms (SPU, SPT), a pointless registry for
> one word (blank),

We do not define a registry-of-limited-to-one-entry-size, we define 
extensible registry.  At last two tokens will be defined by HTML5, and 
whoever knows how many will be later.  It is pointless to say "pointless".

> and additional tons of MUSTs and
> MUST NOTs dealing with something that doesn't exist.
>
> This draft is now in the worst state for this year,
> that is not very encouraging for this trivial task.
>
> Claiming that there are no about: IRIs makes no sense
> as soon as there is an<about-query>, cf. RFC 3987
> <iquery>  and<ifragment>.

I know absolutely no application supporting 'about' IRIs.  Please 
provide some examples of 'about' IRIs which are in use now, or which 
make use of i18ned queries/fragments.

Mykyta Yevstifeyev

>
> -Frank
> _______________________________________________
> apps-discuss mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/apps-discuss
>