OASIS Open Document Format for Office Applications (OpenDocument) TC

Dear TC members,

we are discussing the "table:is-sub-table" attribute for a while now, 
but have not come to a conclusion yet.

What we have agreed is that the table:is-sub-table attribute should not 
be used to simulate merged table cells, and that the ODF specification 
should make that clear. What is still under discussion is the question 
whether it may be reasonable to deprecate the attribute at all.

We tried to get some input to the discussion by checking use cases for 
the is-sub-table attribute. One use case that was mentioned are sudoku 
tables, but there seem to be different opinions whether this is a good 
example or not. Another use case that was mentioned were spreadsheet 
tables, where sub tables in combination with the is-sub-table attribute 
could be used to store the result of array functions. However, 
sub-tables are not used that way as of today by the implementations that 
are known to us.

So, taking it all together, it seems to me that we cannot clearly say 
whether it is reasonable to deprecate the attribute or not based on use 
cases.

We discussed the situation in the last work call, and what we observed 
where was that a file format specification like ODF defines features, 
but cannot guarantee that these are used only in reasonable ways. This 
applies to the is-sub-table attribute, but to any other features as 
well. The reason we are discussing the attribute is that the ODF 
specification itself suggested to use it instead of col- and row-span 
attribute, which turned out to be inappropriate. We already agreed to 
change that. So based on that, the following recommendation has bee 
worked out (see
http://lists.oasis-open.org/archives/office/200707/msg00069.html):

a) don't imply in the main specification that subtables should be used
to structure tables
b) note in A11Y guidelines that subtables should not be used and that
col/row-span be used instead...

The a11y guidelines contain already recommendations regarding 
is-sub-table. In the main specification, the examples in section 8.2.6 
would have to be removed, since they are the only hint that is-sub-table 
may be used as a replacement for row- and col-span.

Best regards

Michael


-- 
Michael Brauer, Technical Architect Software Engineering
StarOffice/OpenOffice.org
Sun Microsystems GmbH             Nagelsweg 55
D-20097 Hamburg, Germany          michael.brauer@sun.com
http://sun.com/staroffice         +49 40 23646 500
http://blogs.sun.com/GullFOSS

Sitz der Gesellschaft: Sun Microsystems GmbH, Sonnenallee 1,
        D-85551 Kirchheim-Heimstetten
Amtsgericht Muenchen: HRB 161028
Geschaeftsfuehrer: Wolfgang Engels, Dr. Roland Boemer
Vorsitzender des Aufsichtsrates: Martin Haering

On Thursday 26 July 2007 12:25:37 Michael Brauer wrote:
> So, taking it all together, it seems to me that we cannot clearly say
> whether it is reasonable to deprecate the attribute or not based on use
> cases.

Hmm; we can't come up with a usecase that clearly shows the advantages of 
the feature, and removing the feature does not disable known usecases.

I can only conclude one thing from those two points.

ps. I assume that nested tables are still possible without this feature. 
So we are not talking about removing any real functionality. Right?
-- 
Thomas Zander

Hi Thomas,

Thomas Zander wrote:
> On Thursday 26 July 2007 12:25:37 Michael Brauer wrote:
>> So, taking it all together, it seems to me that we cannot clearly say
>> whether it is reasonable to deprecate the attribute or not based on use
>> cases.
> 
> Hmm; we can't come up with a usecase that clearly shows the advantages of 
> the feature, and removing the feature does not disable known usecases.

Well, it is not that we did not find usecases. There only seem to be 
different evaluations of them.

We further have to take into account that we are not talking about 
adding a feature, but about deprecating a feature that we have in the 
specification for more than two years now. And that even the (wrong) use 
case of merging cells was just an example that we provided. We don't 
know whether there are implementations and/or documents that use that 
feature already for a reasonable purpose. We only know that we did not 
find examples where we unanimously agree that they are good examples.

Honestly, only based on the use cases we found I cannot clearly say 
whether I'm personally in favor of deprecating the is-sub-table 
attribute or not. But in this situation, I think it is a reasonable 
approach to correct exactly the issue that has been reported to us, and 
that's to my understanding the use of the is-sub-table attribute for 
merging cells.

> 
> I can only conclude one thing from those two points.
> 
> ps. I assume that nested tables are still possible without this feature. 

Yes.

> So we are not talking about removing any real functionality. Right?

No. Tables that have the is-sub-table attribute set behave differently 
than tables that don't have it set. One difference are the borders. 
Another difference is that tables that have the is-sub-table attribute 
set fill the entire height of the surrounding cell. This means that the 
last table-row gets a different height (it always fills the remaining 
space), what again may have an impact on text that appears vertically 
aligned in that row. Furthermore, the cell addresses for sub tables are 
different than the ones for nested tables.

Best regards

Michael


-- 
Michael Brauer, Technical Architect Software Engineering
StarOffice/OpenOffice.org
Sun Microsystems GmbH             Nagelsweg 55
D-20097 Hamburg, Germany          michael.brauer@sun.com
http://sun.com/staroffice         +49 40 23646 500
http://blogs.sun.com/GullFOSS

Sitz der Gesellschaft: Sun Microsystems GmbH, Sonnenallee 1,
        D-85551 Kirchheim-Heimstetten
Amtsgericht Muenchen: HRB 161028
Geschaeftsfuehrer: Wolfgang Engels, Dr. Roland Boemer
Vorsitzender des Aufsichtsrates: Martin Haering