@Section
    @Title { Spanning columns and rows }
    @Tag { tbl_span }
@Begin
@PP
To make a cell span across several columns, precede the @Code "@Cell"
spanning.columns @Index { spanning columns and rows in tables }
symbol with @Code "@StartHSpan" and replace each spanned cell's
@Code "@Cell" symbol with {@Code "@HSpan"}, like this:
@ID @OneRow @Code @Verbatim {
@Tbl
    rule { yes }
    aformat { @StartHSpan @Cell indent { ctr } @B A | @HSpan | @HSpan }
    bformat { @Cell A | @Cell B | @Cell C }
{
@Rowa
    A { Some famous authors }
@Rowb
    A { Austen }
    B { Chaucer }
    C { Donne }
@Rowb
    A { Balzac }
    B { Darwin }
    C { Goethe }
}
}
The result of this is
@CD @OneRow @Tbl
    rule { yes }
    aformat { @StartHSpan @Cell indent { ctr } @B A | @HSpan | @HSpan }
    bformat { @Cell A | @Cell B | @Cell C }
{
@Rowa
    A { Some famous authors }
@Rowb
    A { Austen }
    B { Chaucer }
    C { Donne }
@Rowb
    A { Balzac }
    B { Darwin }
    C { Goethe }
}
We've used a sample of options to show how naturally these go with
spanning cells:  they apply to the whole cell as usual, whatever
its extent.  It is quite acceptable to span just some of the columns,
not all of them; indeed, there may be no @Code "@HSpan" symbols at
all, and then the cell just spans its own column, which sounds redundant
but actually has a use (Section {@NumberOf tbl_alig}).
@PP
Spanning rows work in the same way; the spanning cell is preceded by
{@Code "@StartVSpan"}, and the spanned cells are replaced by
{@Code "@VSpan"}:
@ID @OneRow @Code @Verbatim {
@Tbl
    rule { yes }
    aformat { @StartVSpan @Cell @I A | @Cell B | @Cell C }
    bformat { @VSpan | @Cell B | @Cell C }
{
@Rowa
    A { Mathematics }
    B { MATH 1001 }
    C { Differential Calculus }
@Rowb
    B { MATH 1002 }
    C { Linear Algebra }
@Rowa
    A { Computer Science }
    B { COMP 1001 }
    C { Introductory Programming }
@Rowb
    B { COMP 1002 }
    C { Introductory Computer Science }
}
}
The result of this is
@CD @OneRow @Tbl
    rule { yes }
    aformat { @StartVSpan @Cell @I A | @Cell B | @Cell C }
    bformat { @VSpan | @Cell B | @Cell C }
{
@Rowa
    A { Mathematics }
    B { MATH 1001 }
    C { Differential Calculus }
@Rowb
    B { MATH 1002 }
    C { Linear Algebra }
@Rowa
    A { Computer Science }
    B { COMP 1001 }
    C { Introductory Programming }
@Rowb
    B { COMP 1002 }
    C { Introductory Computer Science }
}
Here is a notorious larger example, the `spiral':
@ID @OneRow @Code @Verbatim {
@QuotedDisplay @Tbl
    rule { yes }
{
@Row
    format { @StartVSpan @Cell A | @StartHSpan @Cell B | @HSpan }
    A { @SomeText }
    B { @SomeText }
@Row
    format { @VSpan | @Cell B | @StartVSpan @Cell C }
    B { @SomeText }
    C { @SomeText }
@Row
    format { @StartHSpan @Cell A | @HSpan | @VSpan }
    A { @SomeText }
}
}
The @Code "@SomeText" symbol produces a short paragraph of text.  The
result is
@QD @Tbl
    rule { yes }
{
@Row format { @StartVSpan @Cell A | @StartHSpan @Cell B | @HSpan }
    A { @SomeText }
    B { @SomeText }
@Row format { @VSpan | @Cell B | @StartVSpan @Cell C }
    B { @SomeText }
    C { @SomeText }
@Row format { @StartHSpan @Cell A | @HSpan | @VSpan }
    A { @SomeText }
}
It is important when constructing mind-boggling tables like this one
to ensure that every format has exactly the same number of @Code "|"
symbols.  Otherwise the number of columns will differ from row to row.
The names given to the entries ({@Code "A"}, {@Code "B"}, {@Code "C"},
etc.) are quite irrelevant:  having a @Code "@Cell D" in one row and a
@Code "@Cell D" in another does not mean that the cells will appear in
the same column.
# @PP
# There is an asymmetry in the spiral above:  the first column
# occupies slightly more space than the other two.  This arises
# because the left margin of the leftmost column is excluded from the
# calculation of how much space is available.  This anomaly might be
# corrected some day.
@PP
There is a @Code "@StartHVSpan" symbol which combines the effects
of @Code "@StartHSpan" and {@Code "@StartVSpan"}.  You need to
use it in this arrangement:
@ID @OneRow @Tbl
    mv { 0.5vx }
    aformat { @Cell @Code A | @Cell @Code B | @Cell @Code C }
{
@Rowa
    A { "@StartHVSpan" }
    B { "@HSpan" }
    C { "@HSpan" }
@Rowa
    A { "@VSpan" }
@Rowa
    A { "@VSpan" }
}
The blank positions should be left empty.  For example:
@ID @OneRow @Code @Verbatim {
@Tbl
    rule { yes }
    aformat { @Cell A | @Cell B | @Cell C | @Cell D }
    bformat { @Cell A | @StartHVSpan @Cell i { ctr } iv { ctr } B | @HSpan | @Cell D }
    cformat { @Cell A | @VSpan |      | @Cell D }
{
@Rowa
@Rowb
    B { CPU }
@Rowc
@Rowa
}
}
produces
@CD @OneRow @Tbl
    rule { yes }
    strut { no }
    aformat { @Cell A | @Cell B | @Cell C | @Cell D }
    bformat { @Cell A | @StartHVSpan @Cell i { ctr } iv { ctr } B | @HSpan | @Cell D }
    cformat { @Cell A | @VSpan | | @Cell D }
{
@Rowa
@Rowb
    B { CPU }
@Rowc
@Rowa
}
This example illustrates how Lout apportions space in the presence of
spanning columns.  If the spanning cell is naturally narrower than the
cells it spans, it is widened to their size.  If it is wider (as in
the example above), then the last spanned cell is widened to take
up the slack.  This is why the third cell is wider than the second in the
first row of this example.
@End @Section