Add new quote specific milestone attributes - RE: [osis-core] milestone name inconsistencies

[Troy A. Griffitts]

>> I have to strongly disagree with the use of osisID to match up start and
>> end milestones. 
> 
> On what basis? The only reason TEI does not use ID for that purpose is 
> that it has to be unique in the document.

Patrick, sorry, but I agree 100% with Todd on this one.  We don't use 
osisID to match segmented tags either.  We use splitID.  Forcing an 
osisID causes there to be unwanted symbols in the text.

If Harry gets his way when we add the mechanism for defining 
versification schemes, we will USE A REGULAR OSIS DOCUMENT to do such. 
This would mean all kinds of extra anchors, just because we had to add 
them for matching milestones.

We have 2 very different purposes here.  I don't see any convincing 
reason to use the same attribute for both of these purposes.  If you 
want to avoid adding a new attribute, I think splitID is much more 
similar to what we want to do with matching tags.  I would recommend 
using splitID for this purpose or adding a new attribute.

I too thought the old milestone_SE was for matching the start/end.

I still don't get the pt and se.  I think I understand your explanation, 
but have never heard that before, never remember discussing it, and 
really think it is out-of-theme from the rest of the schema.

I would still argue that the mere fact that we HAVE an enumerated list 
of milestone types designates a PROBLEM.  None of these types has any 
special attributes, and I'm sure they should have.  This is the same 
problem we have with <q> and will have with all other "generic 
milestone"-type markup.  If a tag can be used as a milestone, then let 
IT be used as a milestone, or else provide a new tag with all the same 
attributes and behaviour.

<q who="Patrick">in weariness and toil, in sleeplessness often, in 
hunger and thirst, in fastings often, in cold and nakedness -- besides 
all the other things, what comes upon me daily: my deep concern for all 
the osis-core members</q>
Oh wait, that's Paul-- almost.

	-Troy.




> 
> Any mechanism is going to have to involve either pointing, whether 
> expressed as a ref or not, or have some sort of matching tokens.
> 
>>
>> If PT if for presentation type and SE is for Something Else AND neither
>> is to help us match up starting and ending milestones then we need
>> ANOTHER attribute for the purpose of matching up milestones.
>>
>> We must NOT use osisID to match up milestones starting and endings
>> because we would force the inclusion of arbitrary trash in the
>> identifier namespace whose sole purpose is to match up milestones. 
>> If the encoder wants to create an osisID for a milestone let them but it
>> should not be required so that a pair can be matched up.
>>
>> We should have some thing like:
>> <p>
>>     <qStart end="xyz"/> text <qEnd start="xyz"/>
>> </p>
>>
>> If the "end" and "start" attributes are not required what good are start
>> and end milestones.
>>
> What good are the paired milestones anyway? As I recall there were many 
> calls for pseudo-containment milestones, which I personally disagree 
> with but now we have those and paired <qStart>, <qEnd> milestones.
> 
> I am really at a loss to explain why we should require end/start on 
> matching milestones. I have no problem with adding the attributes but 
> see no reason to require their use.
> 
> Waiting on a good reason for requiring end/start attributes (willing to 
> add but why require)?
> 
> Patrick
> 
> 
> 
>>
>>
>> We do not have an attribute to indicate "PresentationType" in any other
>> element and it could be determined to be useful there as well.  I think
>> we should remove milestonePT and force people to use type and subtype to
>> describe the type of data the element is and figure out the
>> "Presentation Type" their style sheet based on the data type. 
>> I still don't know what milestoneSE is for.  Until this morning I
>> thought it was for matching up the "S"tart and "E"nd of a pair of
>> milestones.  If no one can explain what it is for we should remove it.
>>
>> Todd
>>
>>> -----Original Message-----
>>> From: owner-osis-core@bibletechnologieswg.org [mailto:owner-osis-
>>> core@bibletechnologieswg.org] On Behalf Of Patrick Durusau
>>> Sent: Friday, November 15, 2002 12:42 PM
>>> To: osis-core@bibletechnologieswg.org
>>> Subject: Re: Add new quote specific milestone attributes - RE:
>>>
>> [osis-core]
>>
>>> milestone name inconsistencies
>>>
>>> Todd,
>>>
>>> Todd Tillinghast wrote:
>>>
>>>> Patrick,
>>>>
>>>> <snip>
>>>>
>>>>>> Todd
>>>>>>
>>>>>> Does <qEnd> need globalAttributes?
>>>>>>
>>>>> Yes, linked by osisID and splitID values according to the
>>>>>
>>>> documentation.
>>>>
>>>> I don't see any purpose for splitID for a milestone since a milestone
>>>> can not be segmented.
>>>>
>>>> Are you proposing that we "link" <qEnd> with <qStart> with osisID or
>>>>
>> are
>>
>>>> you refereeing to some other meaning of "link"?
>>>>
>>> Linking with osisID. Present due to globalAttributes.
>>>
>>>
>>>>>> Is there a reason that milestoneSe is required?  What good are
>>>>>>
>> start
>>
>>>> and
>>>>
>>>>>> end milestones without a mechanism to find their partner?
>>>>>>
>>>>> I don't read milestoneSe as being required? Reads use="optional"
>>>>>
>>>> What I was trying to say and didn't was that I think "milestoneSe"
>>>> should be required.  (Sorry of the confusion.)  Do you see any that
>>>> milestoneSe should not be required?
>>>>
>>> Sorry, don't see why it should be required. I could simply want to use
>>> milestoneStart and milestoneEnd without the milestoneSe. (This
>>> attribute, like milestonePt, was added at Bob's request for pointing
>>> purposes.)
>>>
>>> Patrick
>>>
>>>
>>>> <snip>
>>>>
>>>> Todd
>>>>
>>> -- 
>>> Patrick Durusau
>>> Director of Research and Development
>>> Society of Biblical Literature
>>> pdurusau@emory.edu
>>>
>>>
>>
>