RE: [oic] Advisory 00007 - Table Rows / Table Columns

["Dennis E. Hamilton" <dennis.hamilton@acm.org>]

OK, I did a more careful analysis of the ODF 1.0 Schema and specifications.  

(I was confusing groups and the collection of rows into <table:table-rows> elements.)

There is a wording problem in ODF 1.0 and I didn't check how it carries over into ODF 1.2.

THE LIMITATION ON <table:rows> and <table:columns>.

In effect, the situation is as follows:

Any sequence of one or more consecutive <table:table-row> elements can be incorporated in a single <table:table-rows> element, except within a <table:table-header-rows> element.

Any sequence of one or more consecutive <table:table-column> elements can be incorporated in a single <table:table-columns> element, except wthin a <table:table-header-columns> element.

The choice to use <table:rows> and <table:columns> is arbitrary.  They are heavily used in the examples in ODF 1.0 and it is clear that these are to clarify the human comprehension of the examples.  The elements have no attributes and serve no function.

The correct statement is that those elements (and groups) cannot be used within table:table-header-rows/-columns elements.  It actually has nothing to do with whether headers are repeated or not.  

WHO HAS HEADERS WHERE?

There is a serious ambiguity in the schemas for table-columns-and-groups and table-rows-and-groups.  This is not what 00007 is about, but it would certainly deserve an advisory of its own.

For example, the table-columns-and-groups pattern is satisfied by a non-empty sequence consisting of <table:table-column-group>, <table:table-columns>, <table:table-header-columns>, and <table:table-columns> elements.  Likewise, a <table:table-row-group> element has the table-columns-and-groups pattern for its content.  Additional constraints are expressed (badly) in the text.  It is clear, however, that a <table:table-header-columns> does not have to be at the beginning of a table-columns-and-groups pattern and that a <table:table-column-group> can have a <table:table-header-columns> child of its own (as can further-nested <table:table-column-group>s).

Great pains were taken to have the table-columns-no-group runs permit an optional header element in any single place in a run.  It is peculiar, after that, that there can be any number of such runs and interspersed groups.  

The table-rows-and-groups pattern works the same way except for rows instead of columns.

IMPORTANT NOTE: The ODF 1.2 Schema is no different.  For ODF 1.2, it is the schema that is authoritative, not the text.  How grand.

 - Dennis



-----Original Message-----
From: oic@lists.oasis-open.org [mailto:oic@lists.oasis-open.org] On Behalf Of Andreas J. Guelzow
Sent: Thursday, October 06, 2011 22:00
To: oic@lists.oasis-open.org
Subject: Re: [oic] Advisory 00007 - Table Rows / Table Columns

On Thu, 2011-10-06 at 22:29 -0600, Dennis E. Hamilton wrote:
> I went back to ODF 1.0 to find out what this is about.  I didn't
> figure out how it works for <table:table-columns> but when I looked at
> <table:table-rows> I notice that these groups can have their own
> headers, and they are definitely considered "grouped."

How do you get that these groups can have their own headers?

In ODF 1.0 in 8.2.4 you have: 

A table must not contain more than one
<table:table-header-rows> element, and a <table:table-rows> must not
follow another <table:table-rows> element. The only exception are tables
that contain grouped rows (see 8.2.5). Such tables contain more than one
<table:table-header-rows> element, provided that they are contained in
different row groups and the rows contained in the elements are
adjacent.

Note that the "grouping" of rows does not happen with table:table-rows
but with table:table-row-group.

> 
> Now, in many arrangements with nested groups, it is desirable to
> repeat the group headers for any group that is split across pages.


Which group headers?

Andreas

> 
> I suspect that there is some sort of intended control over how group
> headings are repeated or not.  There is also an interesting situation
> where most of a group is hidden (leaving a heading and a summary line
> for example).  There could also be a widows-orphans control desired in
> such cases too (i.e., whether to do a page break to keep the rows of a
> group together, or not).
> 
> These are all user-related observations.  Whether and how this sort of
> thing was meant to be incorporated in how groups are paginated is not
> evident in an initial pass on the material -- too much tacit
> knowledge, not enough specification.  
> 
> I don't know if there is any way to check against an implementation
> for confirmation of any of this.
> 
>  - Dennis E. Hamilton

-- 
Andreas J. Guelzow, PhD, FTICA
Concordia University College of Alberta


---------------------------------------------------------------------
To unsubscribe, e-mail: oic-unsubscribe@lists.oasis-open.org
For additional commands, e-mail: oic-help@lists.oasis-open.org