While fixing things related to #87, I noticed a few things that are strange in the current NL sample in the fork:
-
The plain text TEI files have <s> elements in them; this is an anomalous situation, as sentences are (or should be) computed when making the .ana version of the files, the plain text TEI files are assumed to contain text directly inside paragraphs. We cannot check this with XML validation, as the XML schema is the same for plain text TEI and for .ana.TEI files.
-
The <langUsage> element of the corpus root is
<langUsage>
<language ident="nl" xml:lang="en">Dutch</language>
<language ident="nld" xml:lang="en">Dutch</language>
<language ident="sl" xml:lang="en">Slovenian</language>
<language ident="en" xml:lang="en">English</language>
</langUsage>
- you don't need Slovenian (esp. as you have the corpus root marked as
<teiCorpus xmlns="http://www.tei-c.org/ns/1.0" xml:id="PressMint-NL" xml:lang="sl"> which is, of course, wrong, and the validation would comlplain if Slovenian were not defined)
- you have two codes for Dutch, which is wrong, you should have (and use) just the two-letter code
<p> elements have no IDs - this is not strictly speaking and error, but the plain text .txt files will have an empty first column, i.e. it won't be possible to link the plain text to the TEI version.
While fixing things related to #87, I noticed a few things that are strange in the current NL sample in the fork:
The plain text TEI files have
<s>elements in them; this is an anomalous situation, as sentences are (or should be) computed when making the .ana version of the files, the plain text TEI files are assumed to contain text directly inside paragraphs. We cannot check this with XML validation, as the XML schema is the same for plain text TEI and for .ana.TEI files.The
<langUsage>element of the corpus root is<teiCorpus xmlns="http://www.tei-c.org/ns/1.0" xml:id="PressMint-NL" xml:lang="sl">which is, of course, wrong, and the validation would comlplain if Slovenian were not defined)<p>elements have no IDs - this is not strictly speaking and error, but the plain text .txt files will have an empty first column, i.e. it won't be possible to link the plain text to the TEI version.