Re: ISSUE-60: Name of draft should be changed to refer to URI's not URL's

Hi Noah,

On Apr 2, 2009, at 7:08 PM, ext noah_mendelsohn@us.ibm.com wrote:

> John: on today's call, I promised a bit of a followup.

Thank you for doing so.

>
>
> John Kemp wrote:
>
>> What would a '#' mean in a URN? RFC2141[1] suggests that '#' is a
>> reserved character, and would thus
>> require escaping.
>
>
> I'm not quite sure why URNs are coming up as a big consideration  
> wrt/ this
> change.  RFC 3986 [1] is the syntax for all Web identifiers,  
> including for
> example those using the http scheme.  So, the main reason that some  
> of us
> pushed to change URL to URI in the title and content of the draft is  
> that
> it's the preferred initialism for the identifiers we're discussing,
> including those that use http.

Yes, I understand that. As I noted in my email review [1] of the  
document, I even suggested myself that this change be made.

[...]

>
> [1] http://www.ietf.org/rfc/rfc3986.txt

I am not arguing with RFC 3986. I understand that document to be quite  
emphatic in its guidance.

My citing of 2141 was somewhat of a red-herring, but reading that  
document is what caused me to actually think a bit about the change  
from URL to URI.

My concern is about the draft of 'Hash in URIs' [2], which says,  
taking just a couple of examples:

'Designers of URIs have traditionally used ?  to encode server-side   
parameters' - what do 'server-side parameters' actually mean in the  
context of URNs (or non-HTTP URIs even)? Do they have meaning always?  
Does the document describe uses _beyond_  the use of ? as HTTP query  
parameters?

and,

'At its inception, the Web also introduced fragment identifiers  
(preceded by # ) as a means of addressing specific locations in a  
document.' - again, how does this apply in situations where a URI does  
not specify a retrieval algorithm?

What does:

'Create URIs for intermediate pages in a Web application so that the  
back button does the right thing' mean, when the client is not a Web  
browser (or even an HTTP user-agent)?

All of the examples given appear to use http: URIs. My sense is that  
this draft is currently talking mostly, and perhaps exclusively, about  
http: URIs, usually accessed within a Web browser context.

My suggestion is simply that the document ought to say exactly that (I  
think it's useful even if it does only talk about http: URIs), unless  
relevant use-cases beyond http: URIs can be elaborated in the  
document, and the language changed appropriately.

Regards,

- johnk

[1] http://lists.w3.org/Archives/Public/www-tag/2009Mar/0144.html
[2] http://www.w3.org/2001/tag/doc/hash-in-uri.html

Received on Friday, 3 April 2009 12:50:26 UTC