There are two issues:
1) A script should be able to easily extract information from the table. I
have worked with the "Cite journal" template of Wikipedia (that is not
storing a table). It is relatively easy to extract information from this
template by simple regular expressions in Perl. If the information is
stored "just in a table" I guess it would be more difficult due to table
formatting encoding? A template may be better to separate format and
content.
I can't see table syntax being any more difficult to parse than template syntax.
2) An instance of the table may have fields that are
not defined. In my
DTD XML-file defining the table I have presently 33 fields:
<!ELEMENT Loc (activatedSubjects | brainMapExpId | brainMapLocId |
brainMapPaperId | brodmann | comment | coord |
...
valueStd | volume | volumeStd | woroi | x | xReported | xStdReported |
y | yReported | yStdReported | z | zReported | zStdReported |
zScore)*>
A field such as 'brodmann' may or may not be defined for an instance of
the table. Seldom are more than 5-10 fields/columns defined. The table
would be too broad and too empty if all columns/fields are shown.
Ah, now I see the problem. It might be easiest to write a parser hook
to handle it all.