Rashmi Rubdi wrote:
>
> I guess invoking a standalone stylesheet on a named template within it
> makes sense because a styesheet can have multiple templates.
>

There is no necessity for having multiple templates. My own primary
reason for using "-it templatename" is that in our production
environment, we never use the default input, instead some configuration
file is passed on by system property, which in turn contains some
catalog-style references to the sources, which are then read in with
fn:unparsed-text() (we use XSLT for importing structured text, not XML,
which cannot be the principal input).

In the above case I think the named template would be equivalent to
the main() method, because a main() method can be called from the
command line. I am trying to shift my thinking slowly to declarative
programming but sometimes I try to relate with what I already know.

Other usages are a stylesheet that does not need any input (or only some
parameters), i.e., a calculation stylesheet, or the 99-bottles problem,
or a stylesheet that summarizes the online traffic jams by url. Neither
need have multiple primary named templates, but do require an initial
template (well, "require" is too strong a word).

Cheers,
-- Abel Braaksma