Re: [xml-dev] [ANNOUNCE] New MicroXML draft available

[John Cowan <cowan@mercury.ccil.org>]

mozer scripsit:

> 1) there is an extra "[" in the production [3] doctype ::= '<!DOCTYPE'
> s+ name s* [">"

Fixed, thanks.

> 2) I'm not sure I get the side effect of having a special purpose for
> <br/>. Does it mean that a HTML5 browser will be friendly with the
> content of MicroXML document ? If so, at which level ?

It means that a MicroXML document which makes proper use of HTML5
elements and attributes will be valid HTML5.  Unfortunately, <br></br>
is interpreted by HTML5 browsers as <br/><br/>, so by making sure always
to use <br/>, the problem can be avoided.

> 3) xml:space disappeared

It's in Section 5.4.

> 4) it will probably be worth thinking the relationship with XPath Data
> Model, especially for the newly introduced attribute in Processing
> Instruction

One of the main purposes of MicroXML is to avoid the complexity of the
XDM.  Processing instructions don't contain attributes, they just look
as if they do.

> 5) Will there be a MicroSchema specification ? Will it enforce
> attribute on Processing Instruction ?

There are already specs for MicroXSD and MicroRNG.

> 6) There is a asymmetry where you :
> * disallow "<", "&" and ">" in dataChar
> * disallow "<" and "&" in attributeValueChar
> * disallow almost nothing in comment

I wanted to disallow > in attribute values, but James Clark pointed out
that it is allowed (indeed, required) in Canonical XML documents.  This
means that Canonical XML processors can be used to convert full XML to
MicroXML.  I'm open to the possibility of changing this.

Comments are intentionally as open as possible without losing XML
compatibility.

-- 
The experiences of the past show                John Cowan
that there has always been a discrepancy        cowan@ccil.org
between plans and performance.                  http://www.ccil.org/~cowan
        --Emperor Hirohito, August 1945