[osis-core] cell content alignment

[Chris Little]

Patrick Durusau wrote:

> Greetings,
> 
> Chris posted the following on alignment in cells:
> 
>> Right/left justified & centered table cells
>>
>> The contents of table cells, like poetic lines, may be right or left
>> justified or centered.
>>
>> An solution similar to that for <l> is one possibility. 
> 
> 
> His solution for <l> was:
> 
>> <l type="unknown" subType="start|center|end"> for \q1, \qc, & \qr,
>> respectively, assuming LtR text.
> 
> 
> ***Comments on his <l> solution:
> 
> While I have liked a lot of his solutions, this one is the exception. 
> Sorry Chris!

My interest was largely in maintaining symmetry in the mapping of cell & 
line elements that have symmetric elements in USFM.  I'm not especially 
tied to that solution though.

> I don't see "start|center|end" as being subtypes of <l>. For example, I 
> could be marking poetry and have:
> 
> <l type="poetry" subType="2syllables">
> <l type="poetry" subType="4syllables">
> <l type="poetry" subType="6syllables">
> <l type="poetry" subType="8syllables">
> <l type="poetry" subType="2syllables">
> 
> Lest Todd think that is contrived, it is actually a form of poetry 
> invented by the American poet Adelaid Crapsey. Known as Cinquain. (Yes, 
> I cheated and Googled for poetry to find that example. see, 
> http://www.mca.k12.nf.ca/subpro5.htm)

I think [2468]syllables are probably types of lines.  My understanding 
was that we finally decided that <l> implies poetry, and orthographic 
lines (in the CEV or in transcriptions of clay tablets) would be marked 
with <lb/>.

> ****Back to the cell problem****
> 
> Note that <cell> can contain 31 one other elements (see the chart that 
> Steve and I labored so hard to prepare), so <cell> is not the only place 
> where you could record formatting information.
> 
> Depending upon what is inside the <cell> there may be a legitimate 
> element on which to base a particular formating. Or, you may simply want 
> all content of cells left, centered or right aligned. Don't see what 
> putting an attribute on <cell> really buys us.

In my experience, <cell>s usually haven't contained any additional 
elements.  The addition of an element to contain numbers may help 
slightly, since most instances of right aligned cells are due to their 
containing numbers.  (It only helps slightly because it must still be 
determined whether the contents are numerals or words, since the latter 
would not be right-aligned.)

If we want a clear path from USFM to OSIS, I think we need to define a 
mapping to some kind of attribute on <cell>.  If we don't, we don't 
necessarily need to add anything.

> Unless, of course, you think people should be able to left justfy some 
> cells, center others and still others right justify? Seems unlikely at 
> best. I would set the style for the table to do one formatting to all 
> the contents of the cells.

Cell formatting usually occurs uniformally within a column, except for 
the title row (which we already handle).

--Chris