Re: Recognizing Sequences (was More XSL Discussion)

>In this transformation, a simplistic mapping is possible between
><definition-list> and <TABLE>, <term> and <TD>, and <definition> and <TD>.
>There is no simple mapping that results in <TR>.  In an ideal world we
would
>go back to our source DTD and group each <term>/<definition> pair in an
><item> element.  In the real world, this is not possible.  Specifying an
>element sequence that could be mapped into the <TR> is a plausible solution
>that the XSL WG should explore.


The RPG analagy reminds me of how we used to do this in our proprietary 4GL
for processing (and generating reports on) files including heirarchic
(multi-record-type) COBOL files. Essentially the technique was to create a
description of the input file as a Jackson structure (a tree containing
sequence, choice, and iteration nodes: essentially the same as a BNF
definition of the file's syntax) and to associate processing actions with
the nodes of this structure. Much the same as a compiler-compiler, really,
except that the input was a COBOL data file or alternatively a hierarchic
view of a Codasyl database.

The natural conclusion from this line of reasoning might be that we ought
really to be associating processing actions with syntax rules in the
document's DTD. Thoughts?

Mike Kay, ICL


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list