Under Community Review

Just remember, the vote you cast is for the idea, not the proposed name !!! Slight smile 

need for new CALS Inter-Tgroup Rule setting

When you are using multiple Tgroup tables in order to create tables in which the number of columns (or the widths of the columns) change in the middle, you would like to the row rule and row gutter to be consistent when going from tgroup 1 to tgroup with the rest of the table. (= normal row gutter and normal row rule).

With the current settings for the field called "CALS Inter-Tgroup Rule" in the ruling spec that is impossible.
In order to get the normal row rule, you select "rowrule" as the value for that field. (all other possibilities will give you a box rule - which is probably different from the row rule).
When you do so, you get the result as shown:

Yes you get the normal row rule.

But the top gutter of the first row of the second tgroup is set to 0, so the result is not exactly very usable.
However within the (twisted?) XPP logic that is normal. 
The normal row gutter always falls under a row (except for the last row which gets the "below last row gutter value").
The top gutter of a first row of a tables is defined by the value given in the "Above selected box row" field (assuming the box starts above the first row).
So what happens when you have 2 tgroups following each other and you set the "CALS Inter-Tgroup Rule" field to "rowrule"?
Well below the last row of tgroup nr 1 we draw the normal row rule and we get the normal distance between the text and the rule.
And we do not get the gutter value as defined in the "Below Last Row" .
Great!
But at the first row of tgroup 2 things are not so great.
I would expect to see the remainder of the normal row gutter to appear.
Sadly since there is no box rule, we do not get any top gutter at all...

I had opened a ticket on this, but engineering did not want to change the existing behavior as it might upset existing setups.

Hence I would like to have a new value for the "CALS Inter-Tgroup Rule" field that would give me the desired and more logical result.
Maybe we can call this new value "rowruleboth"  or "rowrulenormal" or why not "rowruleBart" Grinning