My answers to https://apps.lis.uiuc.edu/wiki/display/sp09lis590dpl/SummerVacation+XSLT+analysis+assignment are as follows:

1.  That stylesheet would process the "paragraph" and "stress" elements; however, a conforming processor running that stylesheet would also process every other element using the built-in template rules.

2.  That stylesheet will perform work on any well-formed XML document.  The work will be valid HTML (the intended output) only if the stylesheet is being imported by another that is doing quite a lot more and has appropriate input, as the given stylesheet in insufficient to the task regardless of input.

3.  All elements other than "paragraph" and "stress" are explored for "paragraph" and "stress" elements but leave no other evidence in the output.

4.  Whichever template matches an element in the input file first will be used first.  Given the legal usages of "p" and "i" in HTML [spatial ordering] and the possibility that an optimized XSLT processor may not actually be processing the document according to the model--it only has to produce results as if it were--[temporal ordering] there is no way of predicting this with 100% accuracy.

5.  The stylesheet will result in an HTML file if the input is well-formed XML.  Whether or not it's a /valid/ HTML file has already been covered.

6.  "html":  0

7.  "b":  0
    "p":  the number of "paragraph" elements in the input document
    "i":  the number of "stress" elements in the input document

8.
xsl:template match="*|/" mode="#all"
!-- N.B.:  parameters will be inherited throughout this process --
  xsl:apply-templates mode="#current"/
/xsl:template

xsl:template match="text()|@*" mode="#all"
  xsl:value-of select="string(.)"/
/xsl:template

xsl:template match="processing-instruction()|comment()" mode="#all"/

In the absence of other instructions, input elements will explored (see below) and text and attributes will be copied as text; processing instructions and comments will be ignored.  When processing an element, its contents (EXCLUDING attributes) will also be processed.

9.  The question of valid HTML has already been addressed.  The output will be well-formed XML so long as the input document is well-formed XML when stripped of everything but text and "paragraph" and "stress" tags.

10.  You will get a copy of all text in the input file as one long string.

11.  paragraph:p
     stress:i
     !paragraph AND !stress:[null]

12.  Complex mapping (merging or splitting elements, changing node types, etc.), multiple output documents, multiple output formats, context-sensitive mapping (including importing information from elsewhere as part of the transformation), etc.