<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    Dear Dorothee,<br>
    thanks for the pointers. I tried to validate the Akan corpus against
    the schema, but only succeeded after tweaking the PhraseType
    specification in the schema a bit (basically making "globaltags" and
    "word" elements optional).<br>
    <br>
    Then I took a stab at converting it to CLDF, which was fairly easy
    (using your typecraft_python package). The details of this
    conversion are here:
    <a class="moz-txt-link-freetext" href="https://github.com/cldf/cookbook/tree/master/recipes/igt">https://github.com/cldf/cookbook/tree/master/recipes/igt</a><br>
    I think different formats for rather loosely defined things like IGT
    make sense. The idea of CLDF in this respect is to specify only the
    better understood aspects of such datatypes (basically anything that
    can be used automatically) - whereas projects like TypeCraft (or
    XIGT) presumably aim at being able to model and store as much of IGT
    (whatever that means) as possible.<br>
    <br>
    I should note that CLDF also makes it easier to encode metadata in a
    machine-readable way, by piggy-backing on Linked Data: E.g. the
    license information you give for the Akan corpus could be specified
    via<br>
    "dc:license": <a class="moz-txt-link-rfc2396E" href="https://creativecommons.org/licenses/by-nc/4.0/">"https://creativecommons.org/licenses/by-nc/4.0/"</a><br>
    <br>
    Btw.: The DOI (10.13140/RG.2.2.14614.86088) you give for the Akan
    corpus doesn't resolve anymore, but leads here<br>
    <a class="moz-txt-link-freetext" href="https://www.researchgate.net/doi/removed">https://www.researchgate.net/doi/removed</a><br>
    <br>
    best,<br>
    robert<br>
    <br>
    <br>
    <div class="moz-cite-prefix">On 04.04.2018 20:36, Dorothee Beermann
      wrote:<br>
    </div>
    <blockquote type="cite"
      cite="mid:76b213f6-854b-ebce-1e9f-f2c3dee0858c@ntnu.no">
      <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
      <p>Dear Robert,</p>
      <p>Thanks for the feedback. Our <span class="Y0NH2b CLPzrc">XML
          schema definition you find here<b>:</b></span> <a
          class="moz-txt-link-freetext"
          href="https://typecraft.org/typecraft.xsd"
          moz-do-not-send="true">https://typecraft.org/typecraft.xsd <span
            class="Y0NH2b CLPzrc"><br>
          </span></a></p>
      <p><span class="Y0NH2b CLPzrc">We started the development of our
          IGT-XML (TC-XML) in 2006/7,  at that time XIGT was not around
          yet. It was first presented in 2014, as far as I recall. <br>
        </span></p>
      <p>The most common IGT type is the basic three-line interlinear
        format, a format that can also be exported from TypeCraft.  Our
        Akan data is  part of speech tagged in addition.  The TypeCraft
        editor allows for annotations on several tiers which is also
        reflected in our XML.  <br>
      </p>
      <p>I agree with you; its is a good idea to also offer a CSV
        format. We do not do that at the moment, although it is an
        option, since we work with a PostgreSQL database.</p>
      <p>Best,</p>
      <p>Dorothee<br>
      </p>
      <p><br>
      </p>
      <div class="moz-cite-prefix">On 04. april 2018 11:19, Robert
        Forkel wrote:<br>
      </div>
      <blockquote type="cite"
        cite="mid:1afc924d-5b7f-2bef-ccf1-4dc917b349a1@shh.mpg.de"> Dear
        Dorothee,<br>
        I just had a brief look at the Akan corpus. I'd be curious what
        guided your decision to come up with a custom XML based export
        format. The namespace URL <br>
        <pre id="line1"><span><a class="attribute-value" moz-do-not-send="true">http://typecraft.org/typecraft</a></span></pre>
        doesn't seem to resolve, so I guess there is no schema defining
        the XML, right? We included (very basic) support for IGT in CLDF
        (see <a class="moz-txt-link-freetext"
          href="https://github.com/cldf/cldf/tree/master/components/examples"
          moz-do-not-send="true">https://github.com/cldf/cldf/tree/master/components/examples</a>),
        because<br>
        - the examples we found in databases like WALS could be modeled
        in this simplistic form and<br>
        - CSV is better suited for tools like version control than XML<br>
        - we wanted to have IGT data available in the same format
        framework as other linguistic data to make links between data
        homogenous.<br>
        <br>
        We also discussed other IGT formats (see <a
          class="moz-txt-link-freetext"
          href="https://github.com/cldf/cldf/issues/10"
          moz-do-not-send="true">https://github.com/cldf/cldf/issues/10</a>),
        among them XIGT (<a class="moz-txt-link-freetext"
          href="https://github.com/xigt/xigt" moz-do-not-send="true">https://github.com/xigt/xigt</a>),
        which is also an XML format. Did you look at XIGT, and if so,
        why was it not suitable as export format for TypeCraft?<br>
        <br>
        best<br>
        robert<br>
        <br>
        <br>
        <div class="moz-cite-prefix">On 25.03.2018 16:51, Dorothee
          Beermann wrote:<br>
        </div>
        <blockquote type="cite"
          cite="mid:e92280a6-770e-0bb6-c4cd-000f8a36cb7e@ntnu.no">
          <p>Dear all,</p>
          <p>I have followed the discussion on this thread with
            interest. Let me ask you, would any of what you discuss and
            suggest here also apply to Interlinear Glossed Data?<br>
          </p>
          <p>Sebastian talked about making  "typological research more
            replicable". A related issue is reproducible research in
            linguists. I guess a good starting point for whatever we do
            as linguists is to keep things<br>
          </p>
          <div class="moz-forward-container">
            <p>transparent, and to give public access to data
              collections. Especially for languages with little to no
              public resources (except for what one finds in articles),
              this seems essential.<br>
            </p>
            <p>Here is an example of what I have in mind:  We just
              released 41 Interlinear Glossed Texts in Akan. The data
              can be downloaded as XML from:</p>
            <p><a class="moz-txt-link-freetext"
                href="https://typecraft.org/tc2wiki/The_TypeCraft_Akan_Corpus"
                moz-do-not-send="true">https://typecraft.org/tc2wiki/The_TypeCraft_Akan_Corpus</a><br>
            </p>
            The corpus is described on the download page, and also in
            the notes contained in the download. (Note that we can offer
            the material in several other formats.) <br>
            <br>
            <br>
            Dorothee <br>
            <br>
            <font color="#999999" size="-1">Professor Dorothee Beermann,
              PhD<br>
              Norwegian University of Science and Technology (NTNU)<br>
              Dept. of Language and Literature<br>
              Surface mail to: NO-7491 Trondheim, Norway/Norge<br>
              <br>
              Visit: Building 4, level 5, room 4512, Dragvoll,<br>
              E-mail:  <a class="moz-txt-link-abbreviated"
                href="mailto:dorothee.beermann@ntnu.no"
                moz-do-not-send="true">dorothee.beermann@ntnu.no</a><br>
              <br>
              Homepage:<a class="moz-txt-link-freetext"
                href="http://www.ntnu.no/ansatte/dorothee.beermann"
                moz-do-not-send="true">http://www.ntnu.no/ansatte/dorothee.beermann</a><br>
              TypeCraft:<a class="moz-txt-link-freetext"
                href="http://typecraft.org/tc2wiki/User:Dorothee_Beermann"
                moz-do-not-send="true">http://typecraft.org/tc2wiki/User:Dorothee_Beermann</a><br>
            </font><br>
            <br>
            <br>
            <br>
            <br>
            -------- Forwarded Message --------
            <table class="moz-email-headers-table" border="0"
              cellspacing="0" cellpadding="0">
              <tbody>
                <tr>
                  <th nowrap="nowrap" valign="BASELINE" align="RIGHT">Subject:
                  </th>
                  <td>Re: [Lingtyp] Empirical standards in typology:
                    incentives</td>
                </tr>
                <tr>
                  <th nowrap="nowrap" valign="BASELINE" align="RIGHT">Date:
                  </th>
                  <td>Fri, 23 Mar 2018 11:59:18 +1100</td>
                </tr>
                <tr>
                  <th nowrap="nowrap" valign="BASELINE" align="RIGHT">From:
                  </th>
                  <td>Hedvig Skirgård <a class="moz-txt-link-rfc2396E"
                      href="mailto:hedvig.skirgard@gmail.com"
                      moz-do-not-send="true"><hedvig.skirgard@gmail.com></a></td>
                </tr>
                <tr>
                  <th nowrap="nowrap" valign="BASELINE" align="RIGHT">To:
                  </th>
                  <td>Johanna NICHOLS <a class="moz-txt-link-rfc2396E"
                      href="mailto:johanna@berkeley.edu"
                      moz-do-not-send="true"><johanna@berkeley.edu></a></td>
                </tr>
                <tr>
                  <th nowrap="nowrap" valign="BASELINE" align="RIGHT">CC:
                  </th>
                  <td>Linguistic Typology <a
                      class="moz-txt-link-rfc2396E"
                      href="mailto:lingtyp@listserv.linguistlist.org"
                      moz-do-not-send="true"><lingtyp@listserv.linguistlist.org></a></td>
                </tr>
              </tbody>
            </table>
            <br>
            <br>
            <div dir="ltr">Dear all, 
              <div><br>
              </div>
              <div>I think Sebastian's suggestion is very good. </div>
              <div><br>
              </div>
              <div>Is this something LT would consider, Masja?</div>
              <div><br>
              </div>
              <div>Johanna's point is good as well, but it shouldn't
                matter for Sebastian's suggestion as I understand it.
                We're not being asked to submit the coding criteria
                prior to the survey being completed, but only at the
                time of publication. There are initiatives in STEM that
                encourages research teams to submit what they're
                planning to do prior to doing if (to avoid biases), but
                that's not baked into what Sebastian is suggestion, from
                what I can tell.</div>
              <div><br>
              </div>
              <div>I would also add a 4 star category which includes
                inter-coderreliabiity tests, i.e. the original author(s)
                have given different people the same instructions and
                tested how often they do the same thing with the same
                grammar.</div>
              <div><br>
              </div>
              <div>/Hedvig</div>
            </div>
            <div class="gmail_extra"><br clear="all">
              <div>
                <div class="gmail_signature"
                  data-smartmail="gmail_signature">
                  <div dir="ltr">
                    <div>
                      <div dir="ltr">
                        <div dir="ltr">
                          <div dir="ltr">
                            <div dir="ltr">
                              <div dir="ltr">
                                <div dir="ltr">
                                  <div dir="ltr">
                                    <div dir="ltr">
                                      <div dir="ltr">
                                        <div dir="ltr">
                                          <p style="margin:0cm 0cm
                                            0.0001pt;font-size:11pt;font-family:Calibri,sans-serif"><span
                                              style="font-size:9pt"><b><br>
                                              </b></span></p>
                                          <p style="margin:0cm 0cm
                                            0.0001pt"><font face="arial,
                                              helvetica, sans-serif"
                                              size="2"><b>Med vänliga
                                                hälsningar</b><b>,</b><br>
                                            </font></p>
                                          <p style="margin:0cm 0cm
                                            0.0001pt"><b><font
                                                face="arial, helvetica,
                                                sans-serif" size="2">Hedvig
                                                Skirgård</font></b></p>
                                          <p style="margin:0cm 0cm
                                            0.0001pt"><br>
                                          </p>
                                          <p style="margin:0cm 0cm
                                            0.0001pt"><font size="1"><span
style="font-family:verdana,sans-serif;color:rgb(0,0,0)">PhD Candidate</span><br>
                                            </font></p>
                                          <p
style="color:rgb(0,0,0);font-family:Verdana,Helvetica,Arial,sans-serif;margin:0cm
                                            0cm 0.0001pt"><span
                                              style="font-family:verdana,sans-serif"><font
                                                size="1">The Wellsprings
                                                of Linguistic Diversity</font></span></p>
                                          <p
style="color:rgb(0,0,0);font-family:Verdana,Helvetica,Arial,sans-serif;margin:0cm
                                            0cm 0.0001pt"><font
                                              face="verdana, sans-serif"
                                              size="1">ARC Centre of
                                              Excellence for the
                                              Dynamics of Language</font></p>
                                          <p
style="color:rgb(0,0,0);font-family:Verdana,Helvetica,Arial,sans-serif;margin:0cm
                                            0cm 0.0001pt"><font
                                              face="verdana, sans-serif"
                                              size="1">School of
                                              Culture, History and
                                              Language<br>
                                              College of Asia and the
                                              Pacific</font></p>
                                          <p
style="color:rgb(0,0,0);font-family:Verdana,Helvetica,Arial,sans-serif;margin:0cm
                                            0cm 0.0001pt"><font
                                              face="verdana, sans-serif"
                                              size="1">The Australian
                                              National University</font></p>
                                          <p style="margin:0cm 0cm
                                            0.0001pt"><font
                                              color="#666666"
                                              face="arial, helvetica,
                                              sans-serif" size="1"><a
                                                href="https://sites.google.com/site/hedvigskirgard/"
                                                target="_blank"
                                                moz-do-not-send="true">Website</a><br>
                                            </font></p>
                                          <div><br>
                                          </div>
                                          <p style="margin:0cm 0cm
                                            0.0001pt"><br>
                                          </p>
                                        </div>
                                      </div>
                                    </div>
                                  </div>
                                </div>
                              </div>
                            </div>
                          </div>
                        </div>
                      </div>
                    </div>
                  </div>
                </div>
              </div>
              <br>
              <div class="gmail_quote">2018-03-23 0:49 GMT+11:00 Johanna
                NICHOLS <span dir="ltr"><<a
                    href="mailto:johanna@berkeley.edu" target="_blank"
                    moz-do-not-send="true">johanna@berkeley.edu</a>></span>:<br>
                <blockquote class="gmail_quote" style="margin:0 0 0
                  .8ex;border-left:1px #ccc solid;padding-left:1ex">
                  <div dir="ltr">
                    <div>What's in the codebook -- the coding categories
                      and the criteria?  That much is usually in the
                      body of the paper.<br>
                      <br>
                    </div>
                    <div>Also, a minor but I think important point: 
                      Ordinarily the codebook doesn't in fact
                      chronologically precede the spreadsheet.  A draft
                      or early version of it does, and that gets revised
                      many times as you run into new and unexpected
                      things.  (And every previous entry in the
                      spreadsheet gets checked and edited too.)  By the
                      time you've finished your survey the categories
                      and typology can look different from what you
                      started with.  You publish when you're comfortably
                      past the point of diminishing returns.  In most
                      sciences this is bad method, but in linguistics
                      it's common and I'd say normal.  The capacity to
                      handle it needs to be built into the method in
                      advance.  <br>
                    </div>
                    <span class="HOEnZb"><font color="#888888">
                        <div><br>
                        </div>
                        Johanna<br>
                      </font></span></div>
                  <div class="HOEnZb">
                    <div class="h5">
                      <div class="gmail_extra"><br>
                        <div class="gmail_quote">On Thu, Mar 22, 2018 at
                          2:10 PM, Sebastian Nordhoff <span dir="ltr"><<a
href="mailto:sebastian.nordhoff@glottotopia.de" target="_blank"
                              moz-do-not-send="true">sebastian.nordhoff@<wbr>glottotopia.de</a>></span>
                          wrote:<br>
                          <blockquote class="gmail_quote"
                            style="margin:0 0 0 .8ex;border-left:1px
                            #ccc solid;padding-left:1ex">Dear all,<br>
                            taking up a thread from last November, I
                            would like to start a<br>
                            discussion about how to make typological
                            research more replicable, where<br>
                            replicable means "less dependent on the
                            original researcher". This<br>
                            includes coding decisions, tabular data,
                            quantitative analyses etc.<br>
                            <br>
                            Volker Gast wrote (full quote at bottom of
                            mail):<br>
                            > Let's assume that self-annotation
                            cannot be avoided for financial<br>
                            > reasons. What about establishing a
                            standard saying that, for instance,<br>
                            > when you submit a
                            quantitative-typological paper to LT you
                            have to<br>
                            > provide the data in such a way that the
                            coding decisions are made<br>
                            > sufficiently transparent for readers to
                            see if they can go along with<br>
                            > the argument?<br>
                            <br>
                            I see two possibilities for that: Option 1:
                            editors will refuse papers<br>
                            which do not adhere to this standard. That
                            will not work in my view.<br>
                            What might work (Option 2) is a star/badge
                            system. I could imagine the<br>
                            following:<br>
                            <br>
                            - no stars: only standard bibliographical
                            references<br>
                            - *         raw tabular data (spreadsheet)
                            available as a supplement<br>
                            - **        as above, + code book available
                            as a supplement<br>
                            - ***       as above, + computer code in R
                            or similar available<br>
                            <br>
                            For a three-star article, an unrelated
                            researcher could then take the<br>
                            original grammars and the code book and
                            replicate the spreadsheet to see<br>
                            if it matches. They could then run the
                            computer code to see if they<br>
                            arrive at the same results.<br>
                            <br>
                            This will not be practical for every
                            research project, but some might<br>
                            find it easier than others, and, in the long
                            run, it will require good<br>
                            arguments to submit a 0-star (i.e.
                            non-replicable) quantitative article.<br>
                            <br>
                            Any thoughts?<br>
                            Sebastian<br>
                            <br>
                            PS: Note that the codebook would actually
                            chronologically precede the<br>
                            spreadsheet, but I fill that spreadsheets
                            are more easily available than<br>
                            codebooks, so in order to keep the entry
                            barrier low, this order is<br>
                            reversed for the stars.<br>
                            <br>
                          </blockquote>
                        </div>
                      </div>
                    </div>
                  </div>
                  <br>
                </blockquote>
              </div>
              <br>
            </div>
          </div>
          <br>
          <fieldset class="mimeAttachmentHeader"></fieldset>
          <br>
          <pre wrap="">_______________________________________________
Lingtyp mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Lingtyp@listserv.linguistlist.org" moz-do-not-send="true">Lingtyp@listserv.linguistlist.org</a>
<a class="moz-txt-link-freetext" href="http://listserv.linguistlist.org/mailman/listinfo/lingtyp" moz-do-not-send="true">http://listserv.linguistlist.org/mailman/listinfo/lingtyp</a>
</pre>
        </blockquote>
        <br>
        <br>
        <fieldset class="mimeAttachmentHeader"></fieldset>
        <br>
        <pre wrap="">_______________________________________________
Lingtyp mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Lingtyp@listserv.linguistlist.org" moz-do-not-send="true">Lingtyp@listserv.linguistlist.org</a>
<a class="moz-txt-link-freetext" href="http://listserv.linguistlist.org/mailman/listinfo/lingtyp" moz-do-not-send="true">http://listserv.linguistlist.org/mailman/listinfo/lingtyp</a>
</pre>
      </blockquote>
      <br>
      <pre class="moz-signature" cols="72">-- 
Professor Dorothee Beermann, PhD
Norwegian University of Science and Technology (NTNU)
Dept. of Language and Literature
Surface mail to: NO-7491 Trondheim, Norway/Norge

Visit: Building 4, level 5, room 4512, Dragvoll,
Tel.:    +47 73 596525
E-mail:  <a class="moz-txt-link-abbreviated" href="mailto:dorothee.beermann@ntnu.no" moz-do-not-send="true">dorothee.beermann@ntnu.no</a>

Homepage:<a class="moz-txt-link-freetext" href="http://www.ntnu.no/ansatte/dorothee.beermann" moz-do-not-send="true">http://www.ntnu.no/ansatte/dorothee.beermann</a>
TypeCraft:<a class="moz-txt-link-freetext" href="http://typecraft.org/tc2wiki/User:Dorothee_Beermann" moz-do-not-send="true">http://typecraft.org/tc2wiki/User:Dorothee_Beermann</a>


</pre>
    </blockquote>
    <br>
  </body>
</html>