DRs related to Part names

MURATA Makoto eb2m-mrt at asahi-net.or.jp
Tue May 21 04:03:21 CEST 2013

Chris and John,

2013/5/21 Chris Rae <Chris.Rae at microsoft.com>:
> Hi all - many thanks for reposting this, Murata-san. As we mentioned on the WG4 call last week, John and I have been looking through this document and have a couple of questions:
> 1. From item 2: As we discussed on the conference call, we think that ISO/IEC 295000 does not intent permitting a part name with a trailing slash, but the non-terminal ipath-absolute in RFC 3987 does. To this end, we think that we should define part name as 1*("/" isegment-nz).

Agreed.  There are no reasons to allow "/" at the end or "//".

> 2. From item 2: The list of constraints is a subset of the original constraints in 9.1.1. Can you remember why the other constraints weren't included?

1) A part URI shall not be empty. [M1.1]

is retained.

2) A part URI shall not have empty segments. [M1.3]

This is required and my wording "An isegment shall be non-empty"
and your definition (i.e., 1*("/" isegment-nz)) capture this.

3)  A part URI shall start with a forward slash (“/”) character. [M1.4]

This is captured by both your definition (i.e., 1*("/" isegment-nz)) and
mine (ipath-absolute).

4) A part URI shall not have a forward slash as the last character. [M1.5]

My wording "An isegment shall be non-empty" implies this constraint.

5) A segment shall not hold any characters other than pchar characters. [M1.6]

This has to changed for allowing non-ASCII characters in
part names.   We have to allow ipchar.  We only have to use
isegment and  isegment-nz.

> 3. From item 2: Are asterisk and colon characters to be prohibited entirely inside part names, or are we prohibiting only part names that are an asterisk or colon in their entirety?

Anywhere in part names.

> 4. From item 12: Please can you specify which exact five steps are removed from the conversion procedure?

They are shown in the word document attached to the mail
available at:


Hope this helps.


More information about the sc34wg4 mailing list