[osis-core] paragraph break defended once again.

[Steven J. DeRose]

I've read through this thread, and I'm not convinced yet that there's 
a real need.

If you're inserting markup to represent a paragraph "break", it seems 
that you must have made the interpretive decision (justified or not) 
that there is a paragraph around there *to* break. In which case,
marking that paragraph seems to me no more controversial than marking 
it's "break".

I can imagine some peripheral cases where this might not hold, but I 
also think I see practical problems that I think outweigh those:

* Two ways will confuse people.

* If you give people two ways, many will move toward the one that 
requires less thought, even if they know there are good reasons to do 
the other -- everybody's in a hurry, but I want to discourage people 
from delivering a "real" text with markup they choose because they 
were in a hurry. I think if we let paragraph breaks in the door, 
we'll see few paragraph containers, and that practice will not 
gradually go away, because it's too easy *in the moment* to do it the 
less useful way.

* Those paragraphs indiicated by markers would not be processable 
easily by current software. For example, you couldn't translate them 
via XSLT, or associate style info with them via XSL or CSS, or 
retrieve them with various markup-aware tools.  This is especially 
nasty since there also won't be any error reported -- just silent 
misbehavior that might get caught by a very careful proofreader 
later. I think this is what lawyers call an "attractive nuisance" -- 
very tempting to fall into, like an unfenced swimming pool.

I could make the same argument for elminiating all the current 
"break" milestones, but I don't think it would have anywhere near the 
force. For example, querying based on whether two things occur in the 
same page or column is way less valuable (frequent) than in the same 
paragraph; and formatting that occurs merely because you're in a 
column or line seems unlikely.

I think Troy has raised a really good and legitimate issue here, and 
it deserves thorough discussion. At the same time, I'm not convinced 
the benefits would outweigh the costs.

S