[osis-core] 4 issues

[Troy A. Griffitts]

> Rational:
> 
> <w morph="robinson:N-NSM">
> 
> "I'm a WORD who's MORPH is a ROBINSON value N-NSM"
> 
> <seg type="x-occurance:157">
> 
> "I'm a SEG who's TYPE is a OCCURANCE value 157

This logic is flawed.  My apologies:

in the first, MORPH = N-NSM further articulated with ROBINSON
in the second, TYPE <> 157

But Chris tells me that you have decide to recommend:

<num value="157" type="cardinal" subType="x-occurrence">This word occurs 
one hundred and fifty seven time in the New Testament</num>


I feel this is flawed in another way:

"This word occurs..." <> Num

It's a greekWordOccuranceStatement.  Which I would never ask for a tag 
for..., so I think <seg type="x-ntOccurance"> defines it much better: 
It is: A SEG of text who's type is x-ntOccurance.

I just want a way to supply misc details to this element.

__________

I think this might be the actual deficiency:  We've catered to allowing 
new pseudo element tags, e.g. <seg type="x-ntOccurance"> <milestone 
type="x-ntOccurance"> etc.... so <seg type="x-... is our mechanism to 
extend the element tag name list.

What we have NOT catered for is the ability to add new attribute values 
(I realize these can all be done by subclassing the schema... but I'm 
talking on the same level as x-, right now...)

Could we add something global like misc for an attribute?
e.g.
<w lemma="1" morph="2" misc="key1:val1 key2:val2 key3:val3">


Then my new syntax could be:

<seg type="x-ntOccurance" misc="ntCount:157">This word appears...</seg>



	-Troy.