Re: Updated URI Template proposal

James M Snell scripsit:

> The spec will likely need to indicate how new op codes are defined.  How
> should unknown codes be handled?  Are applications free to create their
> own op codes?  Is an iana registry needed? etc.

Well, of course no spec can control the behavior of implementations that
don't claim to conform to it.  I think an IANA registry would be the
proper compromise between "Do as thou wilt" and "Never change anything".

> I would think that unicode codepoints would be what folks would
> typically expect.  

+1

> Next, regarding the varname definition,
> 
>   token varname    '[a-zA-Z0-9][a-zA-Z0-9\_\.\-]*' ;
> 
> I'm curious as to why these additional characters should not be allowed
> in a varname?
> 
>   !@#$%^&()\/:;'[]"
> 
> For instance, allowing : and / would allow us to do...
> 
>   {-prefix|foo/|http://example.org/text/username}

I don't understand what application scenario would call for
the use of universal names in templates.  It seems to me
that local names are sufficient (and more readable; it's
much easier to see where the varname stops if it looks
like an identifier).

> Also, there had been mention of an escape/quote character for varnames,

I have the same question:  for what?

-- 
John Cowan                                cowan@ccil.org
I amar prestar aen, han mathon ne nen,    http://www.ccil.org/~cowan
han mathon ne chae, a han noston ne 'wilith.  --Galadriel, LOTR:FOTR