Re: URIs for the standard output and input streams

I think

std:in
std:out
std:err

win on the cost basis ... one might want to define the std scheme as 
extensible by future revision, so that if the future suggests, for 
example, the use of file descriptors in URIs we might later get

std:fd/23


(Much software would already, non-conformantly, assign file:stdin to the 
file named stdin in the current working directory)

Jeremy


Colin Paul Adams wrote:
>>>>>> "Noah" == Noah Slater <nslater@bytesexual.org> writes:
> 
>     Noah> On Sat, Jan 12, 2008 at 06:36:50PM -0800, Tim Bray wrote:
>     >> Well, if you want to go near the notoriously-broken file: URI
>     >> spec, you're a braver man than I am.  -T
> 
>     Noah> Okay, sure. :)
> 
> Also, the "cost" of amending software that handles the file scheme is
> surely much greater than introducing new schemes. All existing
> software would automatically be broken by default.
> 
>     Noah> I am guessing IANA's response to a request for registering a
>     Noah> new URI scheme would be "why not use HTTP?" - so along those
>     Noah> lines why not publish an RDF profile at
>     Noah> http://purl.org/posix/ and use:
> 
>     Noah>   http://purl.org/posix/stdin http://purl.org/posix/stdout
>     Noah> http://purl.org/posix/stderr
> 
> Ditto.

Received on Tuesday, 15 January 2008 11:31:54 UTC