Re: [relax-ng] Two errata for RNG

[Norman Walsh <ndw@nwalsh.com>]

/ "MURATA Makoto (FAMILY Given)" <EB2M-MRT@asahi-net.or.jp> was heard to say:
| I propose these two errata for the RELAX NG specification.  If there are no 
| objections from the RELAX NG TC in two weeks, as the chair, I will declare 
| that they are unanimously agreed.  I will then, as the errata editor,  
| prepare draft errata.
|
| The first one is to mention DSDL Part 2.  The second one is for closure 
| under boolean operations.  This comes from Haruo (XDuce) Hosoya.
|
|> Hosoya-san suggested a restriction for ensuring closure under boolean operations.
|> An infinite-nameclass <attribute> is required to have <text/> as its content.
|> I am not completely happy about this, but have not reached a realistic alternative 
|> restriction.  More about this, see page 6 and pages 22 thru 23 of our manuscript at:
|> 
|> http://www.kurims.kyoto-u.ac.jp/~hahosoya/papers/attelm.ps
|
| I believe that nobody would like to constrain attribute values without 
| constraining attribute names tightly.  Thus, users lose nothing because 
| of this change.
|
| James and I have already incorporated Hosoya-san's suggestion in 
| DSDL Part 2.  Although I do not know if any implementations support this 
| already, I believe that this is a correct thing to do.

It would take me a long time to read and understand that paper well
enough to follow the argumentation on pages 6 and 22 and 23. Can you
provide a short gloss in relatively simple prose? (If the answer is
"no", that's fine :-)

|> 1) RELAX NG as an ISO standard
|> 
|> Since RELAX NG becomes an ISO standard, "This is the definitive 
|> specification of RELAX NG" in the abstract is no longer true.  
|> 
|> I would propose 
|> 
|> 	This specification defines RELAX NG, ...
|> 
|> We should also mention the ISO standard as a non-normative document.
|> 
|> 	ISO/IEC IS 19757-2 also defines RELAX NG.  Its technical content of 
|> 	is identical to this specification, but there 
|> 	are editorial differences to satisfy ISO rules.
|> 	
|> 2) Attributes using infinite name classes shall have <text/>
|> 
|> As we have discussed long time ago,  attributes using infinite name classes 
|> should have <text/>.
|> 
|> FDIS of RELAX NG says:
|> 
|> 	Attributes using infinite name classes shall have text as their value.
|> 	More precisely, an attribute element that has an anyName or nsName
|> 	descendant element shall have a text child element. NOTE 2 This
|> 	restriction is necessary for closure under negation.

Ah, so the erratum is explicitly to change the current text in
http://www.relaxng.org/spec-20011203.html:

  Attributes using infinite name classes must be repeated. More
  precisely, an attribute element that has an anyName or nsName
  descendant element must have a oneOrMore ancestor element.

   Note
   This restriction is necessary for closure under negation.

To:

  Attributes using infinite name classes shall have text as their value.
  More precisely, an attribute element that has an anyName or nsName
  descendant element shall have a text child element.

    Note
    This restriction is necessary for closure under negation.

Is that the erratum?

                                        Be seeing you,
                                          norm

-- 
Norman.Walsh@Sun.COM / XML Standards Architect / Sun Microsystems, Inc.
NOTICE: This email message is for the sole use of the intended
recipient(s) and may contain confidential and privileged information.
Any unauthorized review, use, disclosure or distribution is prohibited.
If you are not the intended recipient, please contact the sender by
reply email and destroy all copies of the original message.

PGP signature