RE: note about XRI resolution and normal forms

["Drummond Reed" <drummond.reed@cordance.net>]

Kevin,

Good catch. I looked at the references and you're right -- section 2.3.1 of
XRI Syntax does require that the "xri://" prefix be appended in the
conversion to URI normal form.

However I'm certain we did not intend to be that strict about a QXRI. The
intention was that while the QXRI must be a valid absolute XRI, it can being
either "fully qualified" (i.e., include the "xri://" prefix) or not.

Since the ability of an absolute XRI to be syntactically valid without
including the fully-qualified "xri://" prefix is VERY important from the
human usability standpoint, it was just an oversight that when we referenced
section 2.3.1 of XRI Syntax, as we forgot that it specified that the
"xri://" prefix had to be added.

I'm cc'ing the XRI TC list to add this as errata to be fixed by the editors
in Working Draft 11. Specifically, we will update the reference in section
4.1.1 to section 2.3.1 of XRI Syntax regarding the conversion to URI normal
form to specify that the "xri://" prefix is OPTIONAL.

Thanks for pointing this out. Please do send any other issues or errata you
find as we want to incorporate them all into Working Draft 11.

=Drummond 

-----Original Message-----
From: Kevin Turner [mailto:kevin@janrain.com] 
Sent: Tuesday, May 30, 2006 4:58 PM
To: Drummond Reed
Subject: note about XRI resolution and normal forms

just a quick note of something I ran in to while writing this resolution
code:

section 4.1.1 of XRI Resolution states that the QXRI must be in
URI-normal form as defined in section 2.3.1 of XRI Syntax.  That section
says if the "xri://" prefix has been omitted, prepend "xri://".  But,
reading more closely, the QXRI consists of the authority string
*without* a leading scheme or double slash.

I think I've got it worked out now, but it was confusing what the right
thing was for a while there.