[% PROCESS header.tt
	pagetitle = "Guidelines for XML file uploads to Stemmaweb"
%]
<h1>Stemmaweb Help - XML upload guidelines</h2>
<div id="docco">
  <h3>Guidelines for TEI parallel segmentation input</h3>
  <div id="tei">
    <p>The Stemmaweb uploader can accept a collated text tradition in TEI parallel-segmentation format that adheres to the following guidelines:</p>
    <ul>
    	<li>The file must have an &lt;msDesc&gt; element in its header which contains a list of all witnesses used in the text in the &lt;listWit&gt; element. Each &lt;witness&gt; element within &lt;listWit&gt; must have an xml:id attribute; this is taken to be the witness sigil, and is expected to appear in the "wit" attribute of the relevant apparatus readings within the text.</li>
    	<li>Individual reading words may be wrapped in &lt;w&gt; tags; if the tag has an xml:id it will be preserved as the reading ID.</li>
    	<li>All text within the main &lt;text&gt; element of the TEI file will be taken to be part of the collation. Paragraph and line divisions are not currently preserved.</li>
    	<li>At the point where any witness appears in the collation, this should be signified with a &lt;witStart&gt; tag within its own apparatus. The first of these will therefore be the first element in any collated text, for example:
    	<pre>
    	&lt;body&gt;
          &lt;p xml:id="am_1_1"&gt;
            &lt;app xml:id="AppStart"&gt;
              &lt;rdg wit="#Jer #K #A #F #B #I #D #J #O #V #X #Y #Z #W"&gt;&lt;witStart/&gt;&lt;/rdg&gt;
            &lt;/app&gt;
          [...]
          &lt;/p&gt;
    	&lt;/body&gt;
    	</pre> Likewise, when a witness text ends this should be noted with a &lt;witEnd&gt; tag. Lacunae within the text may be indicated through successive uses of the &lt;witEnd&gt; and &lt;witStart&gt; tags.</li>
    	<li>Readings as they appeared before scribal corrections may be indicated using the "type" attribute on the &lt;rdg&gt; element, as shown here:
    		<pre>
            &lt;app xml:id="App530"&gt;
              &lt;lem wit="#Jer #K #F #B #I #A #D #J #O #V #X #Y #Z #W"&gt;
                &lt;w xml:id="L414"&gt;զոր&lt;/w&gt;
              &lt;/lem&gt;
              &lt;rdg type="a.c." wit="#B"&gt;&lt;w&gt;զո&lt;/w&gt;&lt;/rdg&gt;
            &lt;/app&gt;
            </pre></li>
        <li>An apparatus entry may contain the &lt;lem&gt; tag as well as the &lt;rdg&gt; tag; these are treated as equivalent for the purpose of creating the graph.</li>
    </ul>
  </div>
  
  <h3>Guidelines for Classical Text Editor export</h3>
  <div id="cte">
    <p>The Stemmaweb uploader can accept a collated text tradition exported from CTE, provided that certain guidelines are observed:</p>
    <ul>
    	<li>Common abbreviations (e.g. 'om.', 'add.', and the like) should be set correctly via the Format -&gt; Document... menu under the General tab. Any other abbreviations or notes are likely to be interpreted as literal readings by the uploader.</li>
    	<li>Avoid notations such as "<em>tr. post</em> word"; this is difficult to parse even for a human and almost impossible for a computer.</li>
    	<li>Be extremely careful with overlapping apparatus entries; if, for example, a pair of entries reads:
    		<pre>
    		dominus deus] tr. A
    		deus] deo A
    		</pre>
    		then you are simultaneously telling CTE that A reads "deus dominus" and "deo dominus", and the export will become confused. Ensure that there is no scope for confusion.</li>
    	<li>Ensure that the critical apparatus (and the apparatus siglorum, if any) is marked as such via the Format -&gt; Notes/Apparatus settings menu.</li>
    	<li>Ensure (via the menu item Options -&gt; Preferences, under the 'XML' tab) that the 'Apparatus export' option is set to "&lt;app&gt; tags".</li>
    </ul>
  </div>
</div>
  
[% PROCESS footer.tt %]