W3C home > Mailing lists > Public > uri@w3.org > September 2009

RE: URI Template experience

From: Eran Hammer-Lahav <eran@hueniverse.com>
Date: Wed, 30 Sep 2009 11:17:27 -0700
To: "Manger, James H" <James.H.Manger@team.telstra.com>, URI <uri@w3.org>
Message-ID: <90C41DD21FB7C64BB94121FBBC2E72343784DF22ED@P3PW5EX1MB01.EX1.SECURESERVER.NET>
I am not sure how desirable it is to allow variable names to contain encoded values but I don't have strong views on which character to use. I just need to know which character will be the 'encode' operator so that the other specs I am working on will be a compliant subset of this proposal.

Right now Roy's proposal uses '+' to indicate encoding is required. '%' is used for arrays. If people feel strongly about '+', I am happy to just change my specs to use the same operator.

As for {var!}, every proposal I have seen puts the operator in front of the variable name, not after it (in which case it looks too much like an not operator).

EHL

> -----Original Message-----
> From: Manger, James H [mailto:James.H.Manger@team.telstra.com]
> Sent: Tuesday, September 29, 2009 7:12 PM
> To: Eran Hammer-Lahav; URI
> Subject: RE: URI Template experience
> 
> Eran,
> 
> I think {var} and {var!} would be a better URI template syntax to
> indicate when %-escaping is or isn't required for reserved characters.
> 
> draft-hammer-discovery uses {%var}. Using '%' as the special indicator
> (as opposed to '!') is perhaps more suggestive of what it means.
> However, variable names can then not include any %-escaping themselves.
> For instance, it only works if variable names never need escaping, or a
> separate escaping mechanism just for variables names is added (yuck).
> <http://tools.ietf.org/rfcmarkup?doc=draft-hammer-discovery#section-6>
> 
> I think it would be easier to understand and less error-prone to
> implement if '%' only ever appears in a template (like it only ever
> appears in a URI or IRI) as part of a %xx escape sequence.
> 
> 
> I like '!' as an indicator that a substituted value is allowed to
> include (unescaped) reserved characters. It suggests caution, warning,
> danger… which seems appropriate when the substituted value can "do
> anything".
> 
> 
> 
> James Manger
> James.H.Manger@team.telstra.com
> Identity and security team — Chief Technology Office — Telstra
> 
> 
> ----------
> From: uri-request@w3.org [mailto:uri-request@w3.org] On Behalf Of Eran
> Hammer-Lahav
> Sent: Sunday, 20 September 2009 7:49 AM
> To: Roy T. Fielding
> Cc: Mark Nottingham; Joe Gregorio; URI
> Subject: RE: URI Template experience
> 
> I am working on two drafts (XRD in OASIS, draft-hammer-discovery in
> IETF) which make use of URI templates. Assuming this proposal will not
> be published as an RFC within 6 months, these drafts will need to
> define their own syntax. The current proposed syntax in XRD is
> extremely limited, and only includes simple variable substitution (no
> lists or arrays) and a single operator ('%') to require encoding of
> reserved characters.
> 
> Any reason why the reserved substitution operator is '+' and not '%'
> (as in percent-encoded)?
> 
> I want to explicitly design it to be a compatible subset of the work
> being proposed here to allow libraries written for this proposal to
> process the more limited XRD template syntax as well. The only thing I
> need to accomplish that is to make sure the encoding operator is the
> same (which right now means changing my proposal to use '+' instead of
> '%').
> 
> Any advice on how to deal with the different timelines for these specs?
> 
> EHL

Received on Wednesday, 30 September 2009 18:17:57 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 13 January 2011 12:15:42 GMT