The current twexter xml document has a root element <twext>. <twext> can contain any number of <para>paragraphs</para>. Each <para>graph can contain any number of <line>s. A <line> consists of <chunk>s. And every chunk has a <text> and a <twxt>.

Here's a simple twext xml document:

<twext>
  <para>
    <line>
      <chunk>
        <text>hello</text>
        <twxt>hey</twxt>
      </chunk>
    </line>
  </para>
</twext>

As you may note, the current structure does not contain any meta data, this will be added in the multilingual twexter.

perfect, thanks

[edit] how do we save data

we must choose where to save twext files:

• database?

+ fast searching in single point of truth resource

+ easy to manage, update system

- traps twext texts in one server (system)

• files?

+ findable anywhere if online.. twexml?

- if editable, may cause versionitis

• both?

could a database communicate/update w/ files in the wild?

[edit] twexml?

"twexml" is unique string with no results in google.. would such a unique string help search find twext files in the wild?

[edit] design

[edit] flexible

atomx = id morphemes by

1. meaning
2. category
   1. language
   2. tag