Re: [apps-discuss] Call for Adoption: draft-kerwin-file-scheme

[Graham Klyne <[e-mail]>]

On 01/01/2015 18:15, John C Klensin wrote:
>
>
> --On Thursday, 01 January, 2015 18:01 +0000 Graham Klyne
> <[e-mail]> wrote:
>
>> ...
>> For me, the real value in moving towards a consensus
>> specification for file: URIs is that it would provide a common
>> file naming framework for software libraries that unify local
>> and web access to resources.  The capability is substantially
>> achievable today based on the available and informal
>> specifications, but there are a number of problematic edge
>> cases (like Windows drive letters) that can cause interop
>> problems.
>
> Of course, the more we move in the direction of what sounds to
> me like a desire to use "file:" to reference resources that may
> be either local or remote ("web access"), potentially with
> multiple web locations, the closer we get to "file:" meeting the
> design criteria of persistence and location-independence that
> characterize URNs.

Hmmm.   I'm not sure I see this.  I think file: URIs are still location 
dependent, but the location may depend on the context of use rather than the URI 
string alone.  (To clarify:  I have used file: for local resources and http: for 
remote resources, so not expecting to use file: for both.)

What I was referring to was writing software that can handle local and remote 
resources through a common interface, using URIs as an over-arching scheme for 
referencing both.  I've worked on applications that using implementation of this 
idea over the past couple of years, so its not a new capability - just not (yet) 
very portable.

#g
--

>
> I'm not commenting right now on whether I think that would be
> good or bad (not sure I even have an opinion given how widely
> things that appear to be URIs with a scheme of "file" are used),
> but these waters are muddy already and I think that, if we
> decide to muddy them further, it should be with our eyes open.
>
>    john
>