[osis-core] Ambiguity regarding <work> and when to use <identifier type="OSIS"> and <referenceSystem> and which is required.

Todd Tillinghast osis-core@bibletechnologieswg.org
Sat, 16 Nov 2002 08:10:51 -0700


Harry, 

>From reading a few of your documents yesterday I noticed the following
things (I would have written last night but all IP contact with the
world was lost but magically is working again this morning):

OSISIDS
1) The values in the osisIDs seem valid.  This largely because you are
defining your own reference system for each work.

2) I the documents I looked at I did not see any cases where you used a
work identifier in an osisID.  If you have a specific document in which
you did this I would be happy and interested to look at it.

3) In order for a reference to be created into your documents there must
be a way to identify which work we are talking about.  (This is where
the ambiguity starts.  The issue is when do we use <identifier
type="OSIS"> and when do we use <referenceSystem> to create the fully
qualified identifier.  I will attempt a full treatment of the topic in
another email.)  None of your documents I looked at have either an
<identifier type="OSIS"> or a <referenceSystem> element within <work>.

For example in calvin_calcom8.xml for osisID="xiv.i.p7" the work alias
"calvin_calcom08" is specified by default by
osisIDWork="calvin_calcom08", but it is not possible to create a fully
qualified name because there is not <identifier
type="OSIS">Identifer.Of.The.Work</identifier> in the <work> element.

The document is technically correct but for the type of document it is,
it should have an <identifier type="OSIS">.  I argued for having
<identifier type="OSIS"> being required but Patrick pointed out that a
pastor might want to simply encode his sermon for his own purposes and
not be bothered with creating <identifier>s that no one would use
because his sermons would never be referenced.  In the case of these
documents, I would think that they would be referenced. 

OSISREFS
All of the cases of <reference> elements I saw were scripture
references, if there is a document with other references let me know and
I will look at them.  I think we may have arrived upon another ambiguity
here again because of the special case of "Bible" being the default
reference work.  

In general the "prefix" to an identifier in an osisRef would have to be
defined as a <work> in the header.  

<work osisWork="kjv">
   <identifier type="OSIS">Bible.KJV</identifier>
</work>
....
<reference osisRef="kjv:Ps.42.2"/>

and the "fully qualified form" would be "Bible.KJV:Ps.42.2".

It would be incorrect to encode the fully qualified form in the osisRef
as in <reference osisRef="Bible.KJV:Ps.42.2"/>.  (Note it would be valid
to use to the same value for the osisWork attribute in <work> as the
value of <identifier type="OSIS">)

However, we have said that the default osisRefWork="Bible" and that in
this special case encoders do not have to define a corresponding <work>
element in the header.  This allows <reference osisID="Ps.42.2"/> to
have the "fully qualified form or "Bible:Ps.42.2".

The sticky thing with your documents is that you have put a "work alias"
in your osisRef values without declaring a matching <work> element in
the header.  But you used "Bible" which is the special case that is not
required to have a <work> element because it is the default for
osisRefWork.  

The trouble I see is that people will be looking to your documents for
guidance of how to encode other things and will assume they can encode
<reference osisRef="Bible.KJV:Ps.42.2"/> because they saw <reference
osisRef="Bible:Ps.42.2"/> in your document with no matching <work>
element in the header.

It would seem better to encode either <reference osisID="Ps.42.2"/> and
not create a <work> element in the header OR <reference
osisID="bible:Ps.42.2"/> and define a work element in the header.

The issue that is still at large (will post more later) is whether to
use the <referenceSystem> and/or <identifier type="OSIS"> to derive the
"fully qualified name" in the various cases where they are present alone
and both present.

Is this consistent with everyone else's understanding?

Comments?

Todd


> -----Original Message-----
> From: owner-osis-core@bibletechnologieswg.org [mailto:owner-osis-
> core@bibletechnologieswg.org] On Behalf Of Harry Plantinga
> Sent: Saturday, November 16, 2002 5:54 AM
> To: osis-core@bibletechnologieswg.org
> Subject: RE: [osis-core] Ambiguity regarding <work> and when to use
> <identifier type="OSIS"> and <referenceSystem> and which is required.
> 
> Todd,
> 
> I have not paid that much attention to referenceSystems as yet.
> I've used osisIDs and osisRefs like Bible:Matt.1.1 and I've
> used osisIDs as I was previously using IDs in ThML -- i.e.
> on almost all elements.
> 
> I have this as an example header:
> 
> <osisText osisIDWork="suso_wisdom">
>   <header>
>     <revisionDesc><date>2002-11-15</date><p>Original OSIS 1.1
> version</p>
>     </revisionDesc>
>     <work osisWork="suso_wisdom">
>       <title>A Little Book of Eternal Wisdom</title>
>       <contributor role="x-Markup">Jon Van Howegen</contributor>
>       <creator subType="ccel" role="aut">suso</creator>
>       <creator subType="file-as" role="aut">Suso, Henry
> (1295-1366)</creator>
>       <creator subType="short-form" role="aut">Henry Suso</creator>
>       <subject subType="ccel">Practical; Miscellaneous; </subject>
>       <subject subType="LCCN">BV5080 .S813</subject>
>       <subject subType="lcsh1">Practical theology</subject>
>       <subject subType="lcsh2">Practical religion. The Christian
> life</subject>
>       <subject subType="lcsh3">Mysticism</subject>
>       <date type="Created" calendarType="ISO">1910-04-14</date>
>       <publisher>Burns, Oates, &amp; Washbourne Ltd.,
London</publisher>
>       <type>Text.Monograph</type>
>       <format subType="IMT">text/xml</format>
>       <identifier subType="URL">/ccel/suso/wisdom.html</identifier>
>       <language>en</language>
>       <rights>Public Domain</rights>
>     </work>
> 
> Then, in the document, I have like this:
> 
> <div divType="x-div1" divTitle="Preface" n="ii" osisID="ii">
>    <title type="x-h1" osisID="ii.p0.1">A Little Book of Eternal Wisdom
> </title>
> 
> Ugly isn't it?
> 
> -Harry
> 
> 
> > -----Original Message-----
> > From: owner-osis-core@bibletechnologieswg.org
> > [mailto:owner-osis-core@bibletechnologieswg.org] On Behalf Of
> > Todd Tillinghast
> > Sent: Friday, November 15, 2002 2:40 PM
> > To: osis-core@bibletechnologieswg.org
> > Subject: RE: [osis-core] Ambiguity regarding <work> and when
> > to use <identifier type="OSIS"> and <referenceSystem> and
> > which is required.
> >
> >
> > Harry,
> >
> > Are any of you documents to the point where we could use a
> > couple of them as examples regarding the issue with
> > <identifier type="OSIS"> and <referenceSystem>?
> >
> > If you have a text that uses values in osisID and osisRef
> > attributes that use both the defaults and other works, then
> > we would not need to construct an example.
> >
> > I attempted to access you OSIS library online but it seems to be
down.
> >
> > Sorry for the possible complication regarding the release of
> > your documents, if I know that you had a deadline of today I
> > would have worked harder to get with Troy earlier this week
> > and bring this beast to the forefront.  (We started the
> > attempt to call each other a week ago an finally connected
yesterday.)
> >
> > I believe that whatever the resolution, that if you ALWAYS
> > include BOTH a <identifier type="OSIS"> and a
> > <referenceSystem> element with IDENTICAL values for ALL works
> > then your encodings will be correct regardless of the final
> > interpretation.  This however presumes that it is correct for
> > the two values to be identical.  Given you position on such
> > things I doubt that it will be a problem.  (Troy, do you agree?)
> >
> > Todd
> >
> > > -----Original Message-----
> > > From: owner-osis-core@bibletechnologieswg.org [mailto:owner-osis-
> > > core@bibletechnologieswg.org] On Behalf Of Todd Tillinghast
> > > Sent: Friday, November 15, 2002 12:10 PM
> > > To: osis-core@bibletechnologieswg.org
> > > Subject: RE: [osis-core] Ambiguity regarding <work> and when to
use
> > > <identifier type="OSIS"> and <referenceSystem> and which is
> > required.
> > >
> > > Yes there is no change needed for the schema, but there could be
> > > implications regarding how you encode the child elements of <work>
> > > elements.
> > >
> > > Todd
> > >
> > > > -----Original Message-----
> > > > From: owner-osis-core@bibletechnologieswg.org
[mailto:owner-osis-
> > > > core@bibletechnologieswg.org] On Behalf Of Troy A. Griffitts
> > > > Sent: Friday, November 15, 2002 12:05 PM
> > > > To: osis-core@bibletechnologieswg.org
> > > > Subject: Re: [osis-core] Ambiguity regarding <work> and
> > when to use
> > > > <identifier type="OSIS"> and <referenceSystem> and which is
> > required.
> > > >
> > > > Concerning Todd's comments...  Our conversation was more
> > concerning
> > > the
> > > > INTERPRETATION of the markup, than syntax.  Don't be
> > concerned about
> > > > syntax changes for the schema.
> > > >
> > > > Concerning:
> > > >
> > > > <identifier type="OSIS">
> > > >
> > > > I thought we had unofficially agree to use things like:
> > > > Bible.KJV.MyEdition.
> > > >
> > > > Or at least keep things in a 'class'-like structure.
> > > >
> > > > Harry, I would think if you want to have hard links to your
docs,
> > you
> > > > are more than welcome to include:
> > > >
> > > > <identifier type="CCEL">
> > > >
> > > > But I wouldn't use type="OSIS" for the URLs that Todd has
pointed
> > out
> > > > that you are using.
> > > >
> > > >
> > > > 	-Troy.
> > > >
> > > >
> > > >
> > > > Todd Tillinghast wrote:
> > > > > All,
> > > > >
> > > > > Troy and I spoke on the phone yesterday afternoon regarding
the
> > use
> > > of
> > > > > the values in <identifier type="OSIS"> and
> > <referenceSystem> with
> > > work.
> > > > > He and I had differing understandings about which are
> > required in
> > > what
> > > > > cases, and how they are used to resolve the full
> > meaning of values
> > > in
> > > > > osisID and osisRef attributes.
> > > > >
> > > > > Troy is composing a set of examples to make clear what seems
not
> > to
> > > be
> > > > > clear, and hopefully for us as a group to come to a common
> > > > > understanding.
> > > > >
> > > > > Harry, two concerns regarding releasing the CD for
> > distribution at
> > > the
> > > > > meeting next week.
> > > > > 1) If the way you have encoded <work> elements is inconsistent
> > with
> > > the
> > > > > standard.  Then there will be a large number of documents in
> > > circulation
> > > > > with encoding that is inconsistent with the standard.
> > (This is of
> > > > > course no reflection on your encoding but upon the
> > ambiguity that
> > > > > currently exists.)
> > > > > 2) I noticed in past encodings that you have put a value like
> > > > > http://www.ccel..... for the value of <identifier
type="OSIS">.
> > > > > Although the standard does not specify the form of this
> > value, it
> > > would
> > > > > leave the door open for the future to encode things without
the
> > > > > http://www.ccel.org and simply create an identifier
> > that does not
> > > imply
> > > > > a transport or website.
> > > > > 3) The use of a reference to
> > > > > http://www.bibletechnologies.net/osisCore.1.1.1.xsd to
> > specify the
> > > > > schema rather than a reference to a local copy.
> > > > >
> > > > > I don't think any of these issues will cause
> > significant effort to
> > > > > correct but will make for examples that we would want people
to
> > > follow.
> > > > >
> > > > > Based on today's email I understand you are under a
> > deadline and I
> > > hope
> > > > > that there is still time to make necessary adjustments.  I
think
> > we
> > > > > would all agree that it would be preferred for the
> > sample CD to be
> > > > > consistent with what the standard.
> > > > >
> > > > > Todd
> > > > >
> > > > >
> > > > >
> > > >
> > >
> >
> >