RE: [dita] Keys and Keyref - questions about URI syntax

["Grosso, Paul" <pgrosso@ptc.com>]

I don't know what characters we want to allow in keys, but
I agree with the commentor that we should be clearer.  In
particular, we should use a specific term from 3986.  See
3986 for the definition of pchar, userinfo, unreserved,
pct-encoded, sub-delims, gen-delims, reserved and pick the
one that defines the character set we want for keys.  If
none of those terms is just what we want--after deciding 
that we really do need to define yet a different set of 
characters--find the closest and then add or subtract the 
characters that differ from what we want in keys.

If someone can tell me what characters we really want to 
allow in keys, I'll try to determine what term we want.

paul

> -----Original Message-----
> From: Robert D Anderson [mailto:robander@us.ibm.com] 
> Sent: Monday, 2009 February 02 15:02
> To: dita
> Subject: [dita] Keys and Keyref - questions about URI syntax
> 
> 
> Hello,
> 
> The keys/keyref topic currently contains this language about 
> characters
> that are valid in a DITA key:
> A key value consists of one or more space separated key 
> names. Key names
> consist of characters that are legal in a URI [RFC 3986] and must not
> contain the "{", "}", "[", "]", "/", "#", "?"or space 
> characters. The case
> of key names is significant.
> 
> I've gotten the following review comment this subject:
> The description of what characters can be used is not clear. RFC 3986
> should be referenced as a reference, and more detail should 
> be added on
> what are the "characters that are legal in a URI". Does this mean URI
> reserved characters as they can be used in a URI but only in 
> the context
> that the spec defines, or does it mean URI unreserved 
> characters, which
> already do not contain some of the characters that the DITA 
> spec excluded?
> 
> Could some of those from the keyref workgroup provide better 
> wording? I do
> not have a deep familiarity with URIs, so I'm wary of 
> accidentally changing
> a meaning that they keyref group agreed on.
> 
> For reference - RFC 3986: http://www.ietf.org/rfc/rfc3986.txt
> 
> Thanks -
> 
> Robert D Anderson
> IBM Authoring Tools Development
> Chief Architect, DITA Open Toolkit
>