Yves Forkl wrote:

  Hi

> I suggest using "=" instead of "eq"

  Mmh, just use eq if you compare an atomic value to an other one, and
= if you test if an atomic value is equal to one of several values,
isn't it?

> As a matter of taste, I prefer putting the comparison
> operator rather inside the function definition than into
> each call:

> <xsl:function name="my:match-p-style" as="xs:boolean">
>    <xsl:param name="node" as="element()"/>
>    <xsl:param name="styles"/>
>    <xsl:value-of select="$node/w:pPr/w:pStyle/@w:val = $styles"/>
> </xsl:function>

  This makes your function more specialized (it just checks if the p
element matches, mine is a more general accessor).  That may be what
you want, and maybe not.

> Has somebody further improvements to suggest, which might
> perhaps reduce the amount of redundant code even more?

  Depending on your needs, you'll may find the following usefull:

    <xsl:function name="my:p-by-style" as="element(w:p)*">
      <xsl:param name="ps"     as="element(w:p)*"/>
      <xsl:param name="styles" as="xs:string*"/>
      <xsl:sequence select="$ps[w:pPr/w:pStyle/@w:val = $styles]"/>
    </xsl:function>

  From a sequence of w:p elements, it returns all those whose the style
property is one of the strings passed as second parameter:

    my:p-by-style(ns0:Body/w:p, ('style-1', 'style-2'))

  You can adapt it for your needs.  For example pass it not a sequence
of w:p but one (maybe optional) ns0:Body:

    <xsl:function name="my:p-by-style" as="element(w:p)*">
      <xsl:param name="body"   as="element(ns0:Body)?"/>
      <xsl:param name="styles" as="xs:string*"/>
      <xsl:sequence select="
          $body/w:p[w:pPr/w:pStyle/@w:val = $styles]"/>
    </xsl:function>

  I found this kind of little tool functions very helpfull when you
work on highly rich structured languages.  For example here, the
'style' property of an paragraph is in w:pPr, then w:pStyle, then
@w:val.  You just think "its style", but you have to remember all the
path, with the right namespaces (a nightmare when your data model uses
eavily inheritance in WXS through a plethora of schema documents).

  But the key point is to well choose the functions you'll write, IMHO.
 Too much functions is not helpfull, that just adds complexity to
complexity.

  Regards,

--drkm





















	

	
		
___________________________________________________________________________ 
Dicouvrez une nouvelle fagon d'obtenir des riponses ` toutes vos questions ! 
Profitez des connaissances, des opinions et des expiriences des internautes sur Yahoo! Questions/Riponses 
http://fr.answers.yahoo.com