[DDI-users] Re:Adding tags to the DDI (Andrew L Arnott)

[Mark R. Diggory]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Achim and Andrew,

We can make other namespaces available for extension, this is the whole
point of what we're attempting in the future with the DDI in terms of
Structural Reform. We can make this as "controlled" or "loose" as
determined necessary. I guess my thought is the following.

1.) As long as the basic contents of the DDI (within the DDI Namespace)
are validatable without concern for tags existing in alternate
namespaces. Then the basic structure of the data is maintained.

2.) As long as extensions such as MathML are used to express some
"added" conditional logic around the DDI content without subtracting
from its inherent structure, then it would probably be acceptable.

3.) However, If tools become dependent on the presence and "validity" of
alternate namespaced content appearing in DDI Documents (Such as MathML
or any other namespaced validatable content)), and such extensibility is
not managed properly, then we loose the ability to share DDI content
across platforms, or encounter a significant loss of information
concerning the structure of the data.

So we have to approach the idea of alternate namespaces and content very
carefully. Opening up the DDI to allow any alternate namespace, in any
possible configuration would be in my book, a very bad decision and
result many issues and difficulties.

With this in mind, we should fully explore where benefits can be gained
by including MathML in the variable sections and come up with an example
of its inclusion in a controlled fashion.

We are addressing these issues of such extensibility in our working
groups, and we hope we can come to an acceptable compromise for all the
parties using the DDI.

- -Mark

Andrew Arnott wrote:
| Achim,
|
| The goal is a kind of skip pattern, that's true.  Specifically what I'm
| looking for is an answer to the problem: If I have a page of questions on
| divorce details, but on a previous page the question "Were you ever
| divorced?" were answered No, then the whole page (or even a small
section of
| the page) on divorce later in the questionnaire ought to be skipped.  By
| skipped, I mean that the questions should never show up on the web
page (or
| grayed out, whatever you wish).
|
| So to me, the most convenient way of supporting this with the DDI is just
| surround those divorce details questions in a varGrp and use the
| conditionals within that varGrp.  It sounds like you are suggesting that
| rather than make up the new DDI tags to use right here, we adopt MathML's
| syntax, or leave DDI extensible so that additional schemas can be used
| within DDI -- perhaps schemas best used for this and similar tasks.  While
| that idea appeals to me, MathML's example that you included simples
| extremely complex and hard to read, compared to the three or four lines of
| code I had as an example.  Maybe though, if DDI left some
extensibility for
| third-party schemas in its next version, we could each use our favorite
| schemas to perform that conditional work.
|
| Of course then we lose the 100% compatibility that the DDI gives us right
| now, which would be a sore loss.
|
| - Andrew
| -----Original Message-----
| From: ddi-users-bounces at icpsr.umich.edu
| [mailto:ddi-users-bounces at icpsr.umich.edu] On Behalf Of Joachim Wackerow
| Sent: Wednesday, January 21, 2004 3:35 AM
| To: Data Documentation Iniative Users Group
| Subject: Re: [DDI-users] Re:Adding tags to the DDI (Andrew L Arnott)
|
| Another comment.
|
| The goal seems to me to describe conditional connections between several
| questions/variables. One use of conditional connections is the "skip
| pattern", an other use is the construction of new variables on the basis
| of existing variables. With a formal description of this conditional
| connections it would be possible for example to generate a SPSS command
| setup with "select if" statements out of a DDI description.
|
| A very flexible way would be to extend the DDI elements varGrp and/or
| var with the possibilities of MathML. Then it would not be necessary to
| build new elements, with the chance of new restrictions. The main
| problem would be, how to extend DDI with other DTDs/Schemas. But the
| same question exists in other areas of the DDI, like bibliographic
| information (MARC).
|
| Achim
|
| Below is Andrew's example in MathML:
|
| Possible rendering in Unicode: { VG8F1| V1=5 ? V2=3 ? V3=2 }
|
| <?xml version="1.0" encoding="UTF-8"?>
| <!DOCTYPE math PUBLIC "-//W3C//DTD MathML 2.0//EN"
|             "http://www.w3.org/Math/DTD/mathml2/mathml2.dtd">
| <math xmlns="http://www.w3.org/1998/Math/MathML">
|    <semantics>
|      <set>
|        <bvar>
|          <ci>VG8F1</ci>
|        </bvar>
|        <condition>
|          <apply>
|            <or/>
|            <reln>
|              <eq/>
|              <ci>V1</ci>
|              <ci>5</ci>
|            </reln>
|            <condition>
|              <apply>
|                <and/>
|                <reln>
|                  <leq/>
|                  <ci>V2</ci>
|                  <ci>3</ci>
|                </reln>
|                <reln>
|                  <eq/>
|                  <ci>V3</ci>
|                  <ci>2</ci>
|                </reln>
|              </apply>
|            </condition>
|          </apply>
|        </condition>
|      </set>
|    </semantics>
| </math>
|
|
|
| _______________________________________________
| DDI-users mailing list
| DDI-users at icpsr.umich.edu
| http://www.icpsr.umich.edu/mailman/listinfo/ddi-users
|
|
|
| _______________________________________________
| DDI-users mailing list
| DDI-users at icpsr.umich.edu
| http://www.icpsr.umich.edu/mailman/listinfo/ddi-users

- --
Mark Diggory
Software Developer
Harvard MIT Data Center
http://www.hmdc.harvard.edu
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFADuI0CWrQegKnDf8RAo7pAJ0chx+znQQkrRe81QRLbpoLRWQ6PACcC6Dh
uazdEfKx45dNcEM5fbeK4K8=
=JWY3
-----END PGP SIGNATURE-----