[osis-core] one more past unrsolved issue - deriving the presentation value from a complex osisID

[Patrick Durusau]

Todd,

I will hopefully reach all the outstanding issues that are pending on 
the list and have some suggested "bug" fixes by early next week.

Note that we keep coming back to presentation issues that are being 
confused with markup. Markup is supposed to encode the structure of a 
text, not its presentation. If we wanted presentation, we could have 
used TeX and while the text would not be easily searchable or useful 
with widely available software, there would be minute control over 
presentation.

Consider the TEV problem:

There is nothing in the current schema that prevents:

<p n="Matt.1.2-6a" osisID="Matt.1.2 Matt.1.3 Matt.1.4 Matt.1.5 Matt.1.6 
Matt.1.6a">

Such that your stylesheet checks for "n=" before using osisID as a 
display value. This allows the generic use of osisID in cases where 
there is only one identifier and we only have to write the stylesheet 
routine once.

If you want to have anchors in the text (not displayed to users or 
displayed if you like), use generic milestones to mark the beginning of 
"verses" in the paragraph to allow users to jump to what would be a 
verse in a regular translation. That would mean that they could either 
go to the paragraph or the "verse" so you would need to give them a 
choice if you are building search software.

I realize there are typos and some actual problems in the current schema 
but it was not meant to conform to either how texts are currently held 
or how they may need to be displayed. I really do think we can solve the 
display problems I have seen thus far, although some will be more 
difficult than others.

Patrick

Todd Tillinghast wrote:

>PROBLEM: It is difficult and in some cases impossible to derive the
>verse identifiers that match those in current printed editions from an
>OSIS encoded document.
>
>Examples:
>The TEV renders the following paragraph with "2-6a" as the verse
>identifier.
>Here are some options for how it could be encoded:
><p osisID="Matt.1.2 Matt.1.3 Matt.1.4 Matt.1.5 Matt.1.6 Matt.1.6a">
><p osisID="Matt.1.2 Matt.1.3 Matt.1.4 Matt.1.5 Matt.1.6 Matt.1.6.a">
><p osisID="Matt.1.2 Matt.1.3 Matt.1.4 Matt.1.5 Matt.1.6a">
><p osisID="Matt.1.2 Matt.1.3 Matt.1.4 Matt.1.5 Matt.1.6a">
><p osisID="Matt.1.2 Matt.1.3 Matt.1.4 Matt.1.5 Matt.1.6a kjv:Matt.1.2
>kjv:Matt.1.3 kjv:Matt.1.4 kjv:Matt.1.5 kjv:1.6">
><p osisID="kjv:Matt.1.2 kjv:Matt.1.3 kjv:Matt.1.4 Matt.1.2 Matt.1.3
>Matt.1.4 Matt.1.5 Matt.1.6a kjv:Matt.1.5 kjv:1.6">
><p osisID="Matt.1.2 Matt.1.3 Matt.1.4 Matt.1.5 Matt.1.6">
><p osisID="Matt.1.6 Matt.1.5 Matt.1.4 Matt.1.3 Matt.1.2 Matt.1.6a">
><p osisID="Matt.1.2 Matt.1.4 Matt.1.3 Matt.1.5 Matt.1.6a">
><p osisID="Matt.1.2 Matt.1.4 Matt.1.3 Matt.1.5 Matt.1.6a">
>
>In some cases it is possible to create an algorithm that would derive
>the desired text, but in others the translators may have been
>inconsistent.  In any case such an algorithm is complicated in XSL.
>
>SOLUTION OPTION 1: Put the display value in as one of the osisID values.
>4a <p osisID="Display.2-6a Matt.1.2 Matt.1.4 Matt.1.3 Matt.1.5
>Matt.1.6a">  OR
>4b <p osisID="Matt.1.2-6a.display Matt.1.2 Matt.1.4 Matt.1.3 Matt.1.5
>Matt.1.6a">
>OR
>4c <p osisID="Matt.1.2-6a Matt.1.2 Matt.1.4 Matt.1.3 Matt.1.5
>Matt.1.6a"> where the convention is that the FIRST value when there is a
>list of osisIDs is used for display.
>  This creates a strange value in the "namespace" but makes it easy to
>render and would be HARMLESS otherwise.  The rendering system can either
>employ a fancy algorithm to derive "2-6a" and/or use the value following
>the "Display." OR with a trailing ".display" when such a value is
>present (Options 4a and 4b).  This does not add a questionable
>attribute, does not "misuse" <seg>, does not bend the meaning of
>milestone, and still provides a mechanism to solve the problem.
>
>SOLUTION OPTION 2: Without changing the schema is the following:
><p osisID="Matt.1.2 Matt.1.3 Matt.1.4 Matt.1.5 Matt.1.6 Matt.1.6a"><seg
>type="verseIdentifier">2-6a</seg>text of the verses</p>
>
>This is a non-standard and would likely result in many users ignoring
>the type attribute and incorrectly applying the text.
>
>SOLUTION OPTION 3:
><p><milestoneStart osisID="Matt.1.2 Matt.1.3 Matt.1.4 Matt.1.5 Matt.1.6
>Matt.1.6a" textIdentifier="2-6a"/>text of the verses</p>
>For this special case use a milestone and ADD a "textIdentifier" value
>to <milestone>
>
>SOLUTION OPTION 4: Allow milestone to be non-empty in the same way that
><reference> is non-empty.  This is a handy solution but is contrary to
>the nature of milestones.
><p><milestoneStart osisID="Matt.1.2 Matt.1.3 Matt.1.4 Matt.1.5 Matt.1.6
>Matt.1.6a">2-6a</milestoneStart>text of the verses</p>
>For this special case use a milestone and ADD a "textIdentifier" value
>to <milestone>
>
>SOLUTION OPTION 5:
>Tough this is a rendering issue.  Suck it up and write algorithms to
>solve this problem.
>
>I prefer the 4c, is there a need to change the schema as in options 3
>and 4?
>
>Is there a "best practice" that we can recommend for general use?
>
>Todd
>

-- 
Patrick Durusau
Director of Research and Development
Society of Biblical Literature
pdurusau@emory.edu