Maybe some of my confusion stems from the fact that I only have a 1.5
version of the User Guide, some parts of which are still to be fleshed out,
whereas the OSIS docs on the web site are mostly 1.1.1, while you think in
2.0 terms and have the advantage of having been at the meetings. So, the
most obvious question would then be: Is there (a draft of) a OSIS 2.0 User
Guide available somewhere?

One thing that I want to get my head around now is the whole issue of
continuing elements. The way I read the User Guide (1.5) and the header of
the osisCore.2.0.1.xsd.xml, these exist because paragraphs and quotations
and the like, may cross chapter and verse boundaries. But really I would
like to make a strong case for leaving chapters and verses atomic (ie. with
proper start and end tags and the content in between). After all, these are
the primary ordering criteria (including their encompassing divisions). XML
should in the first place be about content, and only in second order deal
with presentation (which I realize is important too for the overall purposes
of OSIS). This way, "chapter" can also remain a type of division as before
(or were there other reasons for making it a separate type?). I think making
both organizational elements (chapter, verse, div) *and* presentational
elements (q, lg, etc) milestoneable is rather odd. Or is it just me you
can't follow the argument that book/chapter/verse is always derviable from
div/p? Why isn't it possible to make p milestoneable instead of
div/chapter/verse?

Can someone shed some light on this for us/me uninitiated?

Thanks
Thomas Boehme

There are several views of scripture. Each is equally valid.
Perspective: "The XSEM project has chosen to use the perspective that 
comes from the task of publishing Scripture as the primary perspective in 
the encoding model." Other perspectives are modeled as possible, although 
not always in the principal hierarchy: "The Canonical view [books contain 
chapters that contain verses]; The Outline view [sections contain sections 
that contain paragraphs]; The Literary view [a collection of prose and 
poetry -- here are various types of prose i.e., narrative, letters, speeches];
The Speech/Discourse view [a series of 
nested speech segments, that is, direct and indirect quotes]; The 
Linguistic view [sentences contain phrases that are made of parts of 
speech, such as nouns, verbs and adjectives]; The Source view [a series of 
witnesses and renderings with one chosen as primary]."
As an experiment my son Eric created a milestone view for each element and 
was able to transform from one primary view to another though the 
milestoned view.  Since publication is my principle need, the publishing 
view is best for me.


Jim Albright
704 843-0582
Wycliffe Bible Translators






"Thomas BOEHME" <catb(at)gmx.net>
06/07/2004 08:26 AM
Please respond to osis-user

 
        To:     <osis-user(at)whi.wts.edu>
        cc: 
        Subject:        [osis-user] OSIS 2.0 User Doc


Maybe some of my confusion stems from the fact that I only have a 1.5
version of the User Guide, some parts of which are still to be fleshed 
out,
whereas the OSIS docs on the web site are mostly 1.1.1, while you think in
2.0 terms and have the advantage of having been at the meetings. So, the
most obvious question would then be: Is there (a draft of) a OSIS 2.0 User
Guide available somewhere?

One thing that I want to get my head around now is the whole issue of
continuing elements. The way I read the User Guide (1.5) and the header of
the osisCore.2.0.1.xsd.xml, these exist because paragraphs and quotations
and the like, may cross chapter and verse boundaries. But really I would
like to make a strong case for leaving chapters and verses atomic (ie. 
with
proper start and end tags and the content in between). After all, these 
are
the primary ordering criteria (including their encompassing divisions). 
XML
should in the first place be about content, and only in second order deal
with presentation (which I realize is important too for the overall 
purposes
of OSIS). This way, "chapter" can also remain a type of division as before
(or were there other reasons for making it a separate type?). I think 
making
both organizational elements (chapter, verse, div) *and* presentational
elements (q, lg, etc) milestoneable is rather odd. Or is it just me you
can't follow the argument that book/chapter/verse is always derviable from
div/p? Why isn't it possible to make p milestoneable instead of
div/chapter/verse?

Can someone shed some light on this for us/me uninitiated?

Thanks
Thomas Boehme


[...]

I think I see the point you are trying to make, but don't you agree that
allowing milestones on the organizational/canonical level as well as the
content level is a bit of a mishmash? Exactly for the same reason that all
models can be transformed into each other (thanks to milestones) I would
rather make one view the primary one and go with it. Maybe all I am trying
to say is: let's make the canonical one the primary one. This would leave
developers and users less room for interpretation which I think is a good
idea.

Cheers,
Thomas

----- Original Message ----- 
From: <Jim_Albright(at)wycliffe.org>
To: <osis-user(at)whi.wts.edu>
Sent: 07 June 2004 15:42
Subject: Re: [osis-user] OSIS 2.0 User Doc

[...]
speeches]; The Speech/Discourse view [a series of[...]
osis-user(at)whi.wts.edu.[...]

There can only be one view in focus at a time. If you need to be able to 
have others in focus then possibly OSIS should make 4 or 5 OSIS models so 
everyone can be happy. Transforming from one view to another then would be 
trivial. So why do you favor canonical view. From a translation point of 
view (and possibly from interpretation view) narrowing the viewers focus 
to just a verse may miss the proper meaning of a section.

Jim Albright
704 843-0582
Wycliffe Bible Translators






"Thomas BOEHME" <catb(at)gmx.net>
06/07/2004 10:35 AM
Please respond to osis-user

 
        To:     <osis-user(at)whi.wts.edu>
        cc: 
        Subject:        Re: [osis-user] OSIS 2.0 User Doc


I think I see the point you are trying to make, but don't you agree that
allowing milestones on the organizational/canonical level as well as the
content level is a bit of a mishmash? Exactly for the same reason that all
models can be transformed into each other (thanks to milestones) I would
rather make one view the primary one and go with it. Maybe all I am trying
to say is: let's make the canonical one the primary one. This would leave
developers and users less room for interpretation which I think is a good
idea.

Cheers,
Thomas

----- Original Message ----- 
From: <Jim_Albright(at)wycliffe.org>
To: <osis-user(at)whi.wts.edu>
Sent: 07 June 2004 15:42
Subject: Re: [osis-user] OSIS 2.0 User Doc

[...]
in[...]
although[...]
contain[...]
sections[...]
speeches]; The Speech/Discourse view [a series of[...]
of[...]
and[...]
in[...]
User[...]
of[...]
quotations[...]
deal[...]
before[...]
from[...]
osis-user(at)whi.wts.edu.[...]

[...]

On Mon, 7 Jun 2004, Thomas BOEHME wrote:
[...]

There is a set hierarchy model for all OSIS Bibles.  That hierarchy model 
is book (<div type="book">), section (<div type="section">),
paragraph 
(<p>).  None of these elements are milestonable.

I don't think anyone particularly lamented not using book (<div 
type="book">), chapter (<chapter>), verse (<verse>) as the
primary model 
since this is a fairly hideously flawed model.  I think one of our guiding 
principles was the old maxim that easy/common stuff should be easy and 
hard stuff should be possible.  Bibles in which paragraphs cross 
verse/chapter boundaries and verses cross chapter boundaries would be 
impossible to encode if we had gone with un-milestonable <chapter> and 
<verse> elements.

We respect your desire to keep things easy, and that is why you can still
encode Bibles using only containers for book, section, paragraph, chapter,
and verse.  But we could not limit functionality so that more complex 
Bibles (like most real, modern, print Bibles are) could not be encoded 
correctly.

--Chris

Chris,

Sounds like there is dark veil across my eyes... What exactly is the
difference between (equally applicable to chapters):

<p>
<v>....</v>
<v sID="123"/>...
</p>
<p>
....
<v eID="123"/>
<v>...</v>
<p>

and this one:

<p sID="abc"/>
<v>...</v>
<v>...<p eID="abc"/><p sID="xyz"/>...</v>
<v>...</v>
<p sID="xyz"/>

Something similar actually is in some of the OSIS documents on the website
coded as <milestone type="x-p"/>.

For your info: I have the programmer (= user) perspective on this, not the
creator perspective - if that makes a difference to your explanation. I will
never create a bible, so the argument that I *could* encode my bible this
way if I wanted to doesn't matter to me. I would to deal with *your* (?)
bibles in the best way possible.

The reason why I favor a canonical approach is that for online bibles the
two most common ways of access is through a) scripture reference or through
b) word search. For the latter, clearly the model doesn't matter, but for
the former atomic verses are rather advantageous. For example, in a
parallel/synoptic view, paragraphs are likely to be different for each
translation, but verses (flawed as they are) are what unites them and keeps
the views in sync.

What you describe (correct me if I am wrong) is the generation of printable
output from the model, which is pretty straight forward either way as you
can just generate the output from the model as you stream it (eg. using a
SAX parser), but jumping into the model from an arbitrary point is what I am
concerned about (yes, pretty single-minded on my behalf). Some newer SAX
parsers in Java allow for deferred instantiation of the model and
milestonable verses and chapters ruin this scheme for the most part. So, in
my view, for bible "printers" either way should be OK and renders the same
result, for online access, the canonical version is advantageous which is
why I suggest to make that the primary model.

Cheers,
Thomas


----- Original Message ----- 
From: "Chris Little" <chrislit(at)crosswire.org>
To: <osis-user(at)whi.wts.edu>
Sent: 08 June 2004 09:27
Subject: Re: [osis-user] OSIS 2.0 User Doc

[...]
all[...]
trying[...]
leave[...]
good[...]
osis-user(at)whi.wts.edu.[...]

On Tue, 8 Jun 2004, Thomas BOEHME wrote:
[...]
*snip*

In theory, there's not much difference between assigning priority to 
structural description (like <p>) and the BCV hierarchy.  We considered 
both options extensively and chose the structural option partly because it 
was in use by a large set of our users already, partly because it is based 
on actual content in light of centuries of scholarship rather than 
traditions with demonstrable errors, and partly because everyone agreed 
that BCV can easily be extracted from the structural system, but not 
vice versa.
[...]

This is quite simply bad markup.  Any time an "x-" value is used, that 
means it is not part of the standard.  In this case specifically, we have 
an example of VERY BAD markup.  This is exactly the opposite of how 
paragraphs are supposed to be marked up, and was done in spite of the 
document's encoder having been told that this is bad encoding.  Any 
documents abuse the standard in this manner are unlikely to be useful to 
anyone aside from their own encoders and should certainly never be 
released publicly.
[...]

Why does the content model matter?  You can still extract portions of a 
document by BCV reference, using the osisID attribute on book, chapter, & 
verse containers.
[...]

I use OSIS from an encoder's and a programmer's perspective.  I'm not
involved in print publishing at all, though certainly others are.
Book/section/paragraph works well for me precisely because I can still
always grab a section of a document using references on the osisID if I 
need to, but the content model sees Bibles as documents rather than just 
databases with verses for records.

Not to be discouraging, and not that I have any more power than to make 
suggestions, but I don't think there's any chance that we will change our 
basic content model.  The standard is already fairly mature, and there are 
already many applications and encoded documents--all of which would become 
invalid if such a huge change were made.

--Chris

To add to what Chris said.

First, the main users/proponents of the BCV hierarchy are on the
software side where they are trying to extract only a range of text
based on a scripture reference.  When faced with the reality that if we
allowed a BCV hierarchy they would ALSO have to support the literary
hierarchy the option becomes clear that we are all better off with fewer
options.


I encoded a few texts both ways and in general the BCV hierarchy is much
simpler to encode as milestones because there are much more complex
structures present in the literary hierarchy.  For example <div
type="book"><div
type="section><list><item><p>text<q
type="block">....
but even the more common case of <div type="book"><div
type="section"><p> has a number of children that can cross the verse
and
chapter boundaries.  So faced with a decision the standards body opted
for the simpler hierarchy to take on the milestone form.

Todd
[...]
considered[...]
because it[...]
based[...]
agreed[...]
have[...]
to[...]
bibles[...]
but[...]
each[...]
and[...]
a[...]
chapter, &[...]
as[...]
using[...]
what[...]
SAX[...]
So,[...]
the[...]
which[...]
I[...]
just[...]
make[...]
our[...]
are[...]
become[...]

Even without knowing any of the history, I do realize that <milestone
type="x-p"/> is bad encoding. That is why I said "like"... 

Well, maybe this all comes back to my earlier request for some more complete
sample documents. Maybe I have to see one of those properly encoded bibles
to realize why the content matters more. Having twenty bibles or so myself,
I still fail to understand why the content is less problematic than BCV with
all its "demonstrable errors". I don't think any two of my bibles have the
same headings, render the same things as quotes etc. But I think I will
leave it at that since I don't want to bog down this group with my seemingly
naïve questions.

Let me just make one final remark: doesn't it strike you as odd that with
every example I cited from public OSIS documents (including those from the
official OSIS website) you are telling that it is using bad encoding? In my
view (I deal a lot with XML in my everyday job) that says a lot about the
document *and* the markup.

I hope you don't get me wrong: I appreciate the effort. All I am trying to
add are 2cts programmer (and purist) wisdom. I really would like to help
make OSIS a success.

Cheers,
Thomas

-----Original Message-----
From: Chris Little [mailto:chrislit(at)crosswire.org] 
Sent: 09. June 2004 11:11
To: osis-user(at)whi.wts.edu
Subject: Re: [osis-user] OSIS 2.0 User Doc

On Tue, 8 Jun 2004, Thomas BOEHME wrote:
[...]
*snip*

In theory, there's not much difference between assigning priority to
structural description (like <p>) and the BCV hierarchy.  We considered
both
options extensively and chose the structural option partly because it was in
use by a large set of our users already, partly because it is based on
actual content in light of centuries of scholarship rather than traditions
with demonstrable errors, and partly because everyone agreed that BCV can
easily be extracted from the structural system, but not vice versa.
[...]

This is quite simply bad markup.  Any time an "x-" value is used, that means
it is not part of the standard.  In this case specifically, we have an
example of VERY BAD markup.  This is exactly the opposite of how paragraphs
are supposed to be marked up, and was done in spite of the document's
encoder having been told that this is bad encoding.  Any documents abuse the
standard in this manner are unlikely to be useful to anyone aside from their
own encoders and should certainly never be released publicly.
[...]

Why does the content model matter?  You can still extract portions of a
document by BCV reference, using the osisID attribute on book, chapter, &
verse containers.
[...]
suggest to make that the primary model.

I use OSIS from an encoder's and a programmer's perspective.  I'm not
involved in print publishing at all, though certainly others are.
Book/section/paragraph works well for me precisely because I can still
always grab a section of a document using references on the osisID if I need
to, but the content model sees Bibles as documents rather than just
databases with verses for records.

Not to be discouraging, and not that I have any more power than to make
suggestions, but I don't think there's any chance that we will change our
basic content model.  The standard is already fairly mature, and there are
already many applications and encoded documents--all of which would become
invalid if such a huge change were made.

--Chris




--
To unsubscribe send an email with subject 'unsubscribe' to
osis-user(at)whi.wts.edu.
Please contact klowery(at)whi.wts.edu for questions.
http://127.0.0.1:9673/mailingLists/osis-user/archive/2004/2004-06/1086611210
250/1086771916276