<section>
<title>Host Schema Evaluation</title>
<para>
This evaluates:
<replaceable>
Word ML
</replaceable>
</para>
<para>
I found some of the Microsoft documentation including the
Overview of WordprocessingML and
"New XML Fatures of the Microsoft Office Word 2003 Object Model"
by Peter Vogel.
</para><para>
I also prepared a simple WordML XML file and loaded it into Microsoft
Word successfully:
<programlisting>

&lt;w:wordDocument xmlns:w="http://schemas.microsoft.com/office/word/2003/wordml"  xmlns:wx="http://schemas.microsoft.com/office/word/2003/auxHint" &gt;
&lt;w:body&gt;&lt;wx:sect&gt;&lt;w:p&gt;&lt;w:r&gt;&lt;w:t&gt;aaa&lt;/w:t&gt;&lt;/w:r&gt;&lt;/w:p&gt;&lt;/wx:sect&gt;&lt;/w:body&gt;
&lt;/w:wordDocument&gt;
</programlisting>
</para><para>
By experimentation,
I determined  that the two xml namespace definitions were necessary.
</para>


<qandaset defaultlabel='qanda'>
<qandadiv><title>Object Representation</title>
<qandaentry><question id="Q2.1">
<para>
Does the host schema use a generic structural markup model?
</para></question><answer><para>
The basic WordML document consists of just five  tags:
<itemizedlist>
<listitem><para>wordDocument</para></listitem><listitem><para>body</para></listitem><listitem><para>p</para></listitem><listitem><para>r</para></listitem><listitem><para>t</para></listitem></itemizedlist>
<sgmltag>p</sgmltag> is a paragraph, <sgmltag>r</sgmltag> allow
for grouping text into runs to allow them to be formatted.  And
the text itself is inside <sgmltag>t</sgmltag>
It also supports the <sgmltag>section</sgmltag>.

</para></answer></qandaentry>
<qandaentry><question id="Q2.2"><para>
Does the host schema define a "clause" object?</para></question><answer><para>
No, but a <sgmltag>section</sgmltag> or <sgmltag>sub-section</sgmltag>
could be used for this purpose.
(However, Neville Holmes and myself have found problems with how style
sheets are applied to sections in Microsoft Word.
</para></answer></qandaentry>
<qandaentry><question id="Q2.3"><para>
Does the host schema define a paragraph level object that represents a structural or grammatical paragraph?</para></question><answer><para>
Yes, <sgmltag>p</sgmltag>.
</para></answer></qandaentry>
<qandaentry><question id="Q2.4"><para>
Using the host schema, can the clause equivalent object be inserted at arbitrary levels in the
document hierarchy without transformation?
</para></question><answer><para>
Unclear, as paragraphs don't contain sections.
</para></answer></qandaentry>
<qandaentry><question id="Q2.5"><para>
In the host schema, are  element names and the structure sufficiently flexible
that the clause and paragraph level objects can be used for other legal and 
business documents?
</para></question><answer><para>
<sgmltag>section</sgmltag> and <sgmltag>p</sgmltag> elements
are obviously used by Microsoft Word for many purposes.
</para></answer></qandaentry>
</qandadiv><qandadiv><title>Metadata</title>
<qandaentry><question id="Q3.1"><para>
Does the host schema provide a mechanism to add semantic information
about:
<itemizedlist>
<listitem><para>whole documents</para></listitem><listitem><para>distinct objects,such as clauses, within documents?
</para></listitem></itemizedlist>
</para></question><answer><para>
As mentioned below, there is a <sgmltag>DocumentProperties</sgmltag> element
that provides some information on the "whole document."  
I could find no other ways of adding metadata to other parts of the
document.
</para></answer></qandaentry>
<qandaentry><question id="Q3.2"><para>
If so, is the metadata model for the host schema sufficient for contracts or will it be necessary to extend it?
</para></question><answer><para>
The <sgmltag>DocumentProperties</sgmltag> element includes information
on title, version, and author.  It is unclear to me how to extend this.
</para></answer></qandaentry>
<qandaentry><question id="Q3.3"><para>
Does the host schema allow embedded values to be represented and semantic
information to be added to these values?
</para></question><answer><para>
Microsoft Word XML supports field processing.
It seems unnecessarily complex with <sgmltag>fldData</sgmltag> elements and several <sgmltag>fldChar</sgmltag> interspersed among <sgmltag>r</sgmltag> and <sgmltag>t</sgmltag> elements in a precise but difficult-to-use manner.
</para></answer></qandaentry>
</qandadiv><qandadiv><title>Processing Technologies</title>
<qandaentry><question id="Q4.1"><para>
Does the host schema require use of a particular processing technology?
</para></question><answer><para>
Obviously, WordML is designed to be processed by Microsoft Word.
However, one could easily process a simplified subset of this
using other technologies such as XSLT.  (In googling WordML,
there were a few  web articles on creating
WordML documents using XSLT.)
Altsoft sells a product that converts a variety of formats, in particular,
WordML, into PDF.
</para></answer></qandaentry>
<qandaentry><question id="Q4.2"><para>
Does the design of the host schema preclude use of particular currently available
processing technologies?
</para></question><answer><para>
See above.
</para></answer></qandaentry>
</qandadiv><qandadiv><title>Number of Content Objects</title>
<qandaentry><question id="Q5.1"><para>
Does the host schema permit the numbering of clauses, paragraphs, lists
and other objects to be represented in the markup?
</para></question><answer><para>
Lists can be created by including a <sgmltag>listPr</sgmltag> in
a paragraph.  This allows the user to override the numbering for
a particular list item or define a style for
a list including how its numbered.
</para></answer></qandaentry>
<qandaentry><question id="Q5.2"><para>
Does the host schema provide a mechanism to define the numbering schema applied to the document so
that two applications could apply the same numbering, if desired?
</para></question><answer><para>
See answer for previous question.
</para></answer></qandaentry>
</qandadiv><qandadiv><title>Complete Document Representation</title>
<qandaentry><question id="Q6.1"><para>Using the host schema, will it be possible for the contract author to explicitly represent all parts of the
narrative contact terms or will it be necessary to imply some 
parts?
</para></question><answer><para>
The <sgmltag>subsection</sgmltag> and <sgmltag>section</sgmltag> elements
could be used for this purpose.
</para></answer></qandaentry>
<qandaentry><question id="Q6.2"><para>
Does the host schema represent the relationship between all significant 
components in a way that allow high quality print and web rendition of of contact documents?
</para></question><answer><para>
Obviously, Microsoft Word does provide quality style documents.
Extensive style elements and style sheets are avaiable in WordML.
</para></answer></qandaentry>
</qandadiv><qandadiv><title>Variables Definition</title>
<qandaentry><question id="Q7.1"><para>
Does the host schema include a mechanism for defining variables for embedded data values?
</para></question><answer><para>
See above.
</para></answer></qandaentry>
<qandaentry><question id="Q7.2"><para>If the host schema does not include such a mechanism,
is there any obstacle to adding it?
</para></question><answer><para>
See above.
</para></answer></qandaentry>
</qandadiv><qandadiv><title>Ease of use for authors</title>
<qandaentry><question id="Q8.1"><para>
Based on the following factors is the host schema easy for contract authors to use:
<itemizedlist><listitem><para>Does it require authors to know only a small number of elements (positive factor)?
</para></listitem><listitem><para>
Does it require authors make unnecessary or subtle distinctions that will be applied inconsistently (negative factor?)
</para></listitem><listitem><para>
Does it have a clear logical structure that can be quickly explained to new users ( positive factor)?
</para></listitem><listitem><para>
Does it allow authors to re-locate content objects within a document
hierarchy with minimal or no need for transformation of markup
(positive factor)?
</para></listitem>
</itemizedlist>
</para></question><answer><para>
The use of <sgmltag>sub-section</sgmltag>, <sgmltag>section</sgmltag>,
<sgmltag>paragraph</sgmltag>.  Lists are supported, but as
an add-on to paragraphs, they are relatively hard to use.
</para><para>
As, there is a definite hierarchy, moving text around from hierarchy to
hierarchy does not seem feasible.
</para><para>
Some aspects of the model including styling lists and field variables
seem difficult.
</para></answer></qandaentry>
</qandadiv><qandadiv><title>Schema Syntax</title>
<qandaentry><question id="Q9.1"><para>
Is the host schema a DTD only or can it also be expressed as an XML Schema or other schema type?
</para></question><answer><para>
The schema is an XML Schema with multiple name spaces.
</para></answer></qandaentry>
</qandadiv><qandadiv><title>Adaptability to contracts</title>
<qandaentry><question id="Q10.1"><para>
Does the host schema provide for the complete representation for the distinct structures commonly found in contracts?
</para></question><answer><para>
See answers above.
</para></answer></qandaentry>
<qandaentry><question id="Q10.2"><para>
If not, does the host schema explicitly allow additional distinct structures
to be added?
</para></question><answer><para>
I did not see any extension mechanisms on Microsoft's Web site.
</para></answer></qandaentry>
<qandaentry><question id="Q10.3"><para>Does the host schema allow elements not considered necessary for contracts markup
to be removed without contract documents being incompatible in a disadvantageous way with other documents using the host schema?
</para></question><answer><para>
It is unclear how to remove elements from Microsoft's defined
schema.  However, only five elements are needed for Microsoft Word to read
a document in and display it nicely.
</para></answer></qandaentry>
<qandaentry><question id="Q10.4"><para>
If distinct contract structures are added to the host schema, will this result in
contracts documents being incompatible in a
disadvantageous way with other documents using the host schema?
</para></question><answer><para>
I added an element from an unknown 
schema as a test to the WordML XML.  When I loaded the document
into Microsoft Word, the element displayed with
a purple highlighting.
However, I was able to use Microsoft Word to 
put text inside of this element and have it
saved properly in XML.
</para></answer></qandaentry>
</qandadiv><qandadiv><title>Vendor and Developer Support</title>
<qandaentry><question id="Q11.1"><para>
Is the host schema already in widespread or general use for markup of narrative
documents?
</para></question><answer><para>
This is the representation for Microsoft Word docuemnts.
</para></answer></qandaentry>
<qandaentry><question id="Q11.2"><para>Are the already developed applications that Will make it easy of for organizations to implement the TC's specification based around the host
schema?
</para></question><answer><para>
This is the representation for Microsoft Word docuemnts.
</para></answer></qandaentry>
<qandaentry><question id="Q11.3"><para>Is there any reason to expect that the
host schema will prove any particular advantages in gaining market support.

</para></question><answer><para>
As mentioned in this TC's  Requirements Document, most Contracts
are prepared using Microsoft Word.
</para></answer></qandaentry>
</qandadiv><qandadiv><title>Other Factors</title>
<qandaentry><question id="Q12.1"><para>Does the host schema provide any other advantages for use in the TC's specification?
</para></question><answer><para>
</para></answer></qandaentry>
<qandaentry><question id="Q12.2"><para>
Does the host schema have any other disadvantages that make it undesirable for use in the TC's
specification?
</para></question><answer><para>
Neville  Holmes' article in the 
November 2001 issue of <emphasis>IEEE Computer</emphasis> explained
problems in formatting Microsoft Word documents, particularly when
cutting and pasting from  one document into another.  In reading the
specification for the XML representation for these documents, it is obvious
why Microsoft Word has the problems that it does.
</para></answer></qandaentry>
</qandadiv>
</qandaset>
</section>




