Re: Some XSL Questions

[Norman Walsh <norm@xxxxxxxxxxxxx>]

Sean Mc Grath <digitome@xxxxxx> writes:

> >> 1) How many tree walks?
> >> At the end of section 5 there is an example of using a named mode to achieve
> >> a table of contents built from TITLE elements followed by the content
> >> proper. Does the invocation of a named mode initiate a new walk of the
> >> document tree? Every TITLE element will need to be traversed twice, right?
> >
> [Norman Walsh]
> >A named mode does not initiate a new walk of the tree, but
> ><select-elements> does.
> 
> Do you mean the <select from = ...> element in the current draft?
> If so, I understand. Thanks.

Yes.

BTW, it occured to me after my previous post that I ought to say
that there's nothing that _requires_ <select> to do another tree
walk.  It would be possible to annotate the source grove during
initial construction so that <select> could be done more
efficiently (at least usually).  That's irrelevant for
understanding the submission, but if you're concerned about
speed of execution for a stylesheet that does lots of selects,
it might be worth considering. ;-)

> One more question. In section 7.3 there is this:
> 
>   <paragraph start-indent = "=childNumber(element) * 12pt">
> Is the attribute value literal evaled by EcmaScript as a result of the "="?

Yes.  Attribute values that begin with "=" are evaluated by ECMAScript.

> Is the
> intention that the likes of childNumber(element) will be provided by the DOM?

That's unclear. It will certainly be necessary for the DOM and
XSL groups to coordinate thier efforts in defining what
functions an XSL processor is expected to provide, but the
details are murky.

> (what is "element" in the above?)

"element", in the context of that example, is the current node.
The deviding line between functionality in XSL and functionality
that requires ECMAScript is undefined at this point too. ;-)

--norm






 DSSSList info and archive:  http://www.mulberrytech.com/dsssl/dssslist