Re: [Uri-review] Request for review

["Martin J. Dürst" <[e-mail]>]

Hello Timothy,

On 20/05/2020 00:47, Timothy Mcsweeney wrote:
> Hi Henry,
> I apologize for anything that was misleading as that was certainly not my
> intent. I will separate those two statements. The only similarity I wanted to
> point out was that 'tel' and 'leaptofrogans' use less than all five scheme
> components. Perhaps 'geo:' would have been a better example?
> 
> For the syntax, I wasn't sure exactly how much info was needed. I thought that
> only the scheme and path were required. Maybe I could change the reference to
> [RFC3986] section 2.2? If you think it would be better, should I write it out
> more like this?

In the extreme, only the scheme is needed. "dav:" is an example. But 
without a colon, it's not a scheme.

> path = / path-noscheme ; begins with a non-colon segment
>             / path-rootless ; begins with a segment
>             / path-empty ; zero characters
> 
> path-noscheme = segment-nz-nc *( "/" segment )
> path-rootless = segment-nz *( "/" segment )
> path-empty = 0<pchar>
> 
> segment = *pchar
> segment-nz = 1*pchar
> segment-nz-nc = 1*( unreserved / pct-encoded / sub-delims / "@" )
>                          ; non-zero-length segment without any colon ":"
> 
> pchar =      unreserved / pct-encoded / sub-delims / ":" / "@"

I'm not sure what these parts of the grammar are supposed to do here, 
but you can't just start in the middle of the grammar and claim that you 
get an URI.


> Hi Martin,
> I know at first glance it might look out of place but the #fg34htx part isn't a
> fragment.

By the definitions of RFC 3986, it is a fragment (identifier). This is 
independent of what you want to call it.

> I think the "drop" part will be recognized as the scheme name because
> of its dereferencing.

Please do the following, as an easy experiment:

- Create a simple Web page somewhere, e.g. called base.html,
   and in it, include the following part:
   <a href='drop#fg34htx'>Link to drop URI</a>
- In the same directory, create another Web page, with the file name
   simply being 'drop' (without extension). Way down in that Web page,
   include the following:
   <a id='#fg34htx' name='#fg34htx'>Fragment fg34htx</a>
- Activate the link in the first page, and observe how it goes to the
   fragment in the second page.
[If you set up the pages on the server, you may have to take some care 
that the 'drop' file is really served with an HTML media type; this may
be a bit tricky.]

If my explanations don't help, maybe doing this experiment will show you 
what I mean.

Regards,   Martin.

P.S.: The solution is simple. If you change "drop#fg34htx" to 
"drop:fg34htx", then you actually match the URI production and no longer 
have a fragment id.

>> On May 19, 2020 at 5:40 AM "Henry S. Thompson" < [e-mail] 
>> <mailto:[e-mail]>> wrote:
>>
>>
>> Timothy Mcsweeney writes:
>>
>>> This is a request for a review of the 'drop' URI scheme.  The
>>> draft can be found here
>>> https://datatracker.ietf.org/doc/draft-mcsweeney-drop-scheme/
>> Without commenting on any other aspect of the proposed scheme, and
>> mostly just to save people time, I found the following aspect of the
>> proposal somewhat misleading:
>>
>> "Similar to the previously registered 'tel' [RFC3966] and
>> 'leaptofrogans' [RFC8589] URIs, the 'drop' URI scheme is
>> syntactically correct but does not need to use all 5 of the
>> parse-able components available to it. The 'drop' scheme uses the
>> number sign '#' as a general delimiter as seen in Appendix
>> A. Collected ABNF [RFC3986]. The scheme syntax is as follows:
>>
>> " drop-uri = 'drop#' character string
>>
>> drop # fg34htx
>> \__/ \_/ \_____/
>> | | |
>> <scheme> | <scheme-specific-part>
>> <gen-delim>
>> "
>>
>> I read this as implying that
>>
>> 1) 'tel' and 'leaptofrogans' URIs did not begin "tel:" and
>> "leaptofrogans:";
>> 2) The 3986 ABNF for URIs recognises "drop#fg34htx" as a URI.
>>
>> Neither of these is in fact that case. The two referenced schemes
>> require ':' after the 'scheme' component, and the 'URI' production does
>> _not_ recognise the above example. (The 'URI-reference' production does,
>> but not using the 'scheme' production to cover the "drop" part.)
>>
>> ht
>> -- 
>> Henry S. Thompson, School of Informatics, University of Edinburgh
>> 10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
>> Fax: [phone], e-mail: [e-mail] <mailto:[e-mail]>
>> URL: http://www.ltg.ed.ac.uk/~ht/
>> [mail from me _always_ has a .sig like this -- mail without it is forged spam]
>>
>> The University of Edinburgh is a charitable body, registered in
>> Scotland, with registration number SC005336.
>>
>> _______________________________________________
>> Uri-review mailing list
>> [e-mail] <mailto:[e-mail]>
>> https://www.ietf.org/mailman/listinfo/uri-review
> 
> 
> _______________________________________________
> Uri-review mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/uri-review
> 

-- 
Prof. Dr.sc. Martin J. Dürst
Department of Intelligent Information Technology
College of Science and Engineering
Aoyama Gakuin University
Fuchinobe 5-1-10, Chuo-ku, Sagamihara
252-5258 Japan