<group> should be a member of att.typed

raffazizzi commented 11 years ago

When grouping texts, and especially with nested group hierarchies, it would be helpful to assign a type to a group. If one regards <group> as a means to create hierarchies of texts, just like <div> is a mean to create hierarchies within texts, the same logic of distinguishing hierarchical elements by a type applies.

Original comment by: frederik-elwert

raffazizzi commented 11 years ago

Can you provide a sample use case?

Original comment by: lb42

raffazizzi commented 11 years ago

Yes, sure!

We are trying to convert http://tipitaka.org/romn/ to more structured TEI. Here, we have a nested hierarchy of pitaka (collection, literally: basket), nikaya (volume), book, chapter, subhead, subsubhead. For various reasons, we decided to use "chapter" as a single text, everything under it (subhead, subsubhead) as div, and everything above it as groups.

Thus, we would like to do something along these lines:

:::xml
<TEI xmlns="http://www.tei-c.org/ns/1.0">
  <teiHeader/>
  <text>
    <group type="pitaka">
      <head>Suttapiṭaka</head>
      <group type="nikaya">
        <head>Dīghanikāyo</head>
        <group type="book">
          <head>Sīlakkhandhavaggapāḷi</head>
          <text type="chapter">
            <body>
              <div type="subhead">
              content ...
              </div>
            </body>
          </text>
        </group>
      </group>
    </group>
  </text>
</TEI>

Does this make sense?

Original comment by: frederik-elwert

raffazizzi commented 11 years ago

Thanks for the example. I can see no logical reason why we shouldn't allow @type on <group> in the same way as we do on <text> and <div> . However, I am quite curious about the reasons that led you to treat each chapter as a single text, rather than as a div. To my mind a chapter is an incomplete object, whereas the definition (admittedly rather vague) of a <text> suggests that it is complete in itself. I don't know enough (or anything much!) about this material so I cannot judge how pertinent that distinction is to your work. However, as aforesaid, I see no problem in adding @type to <group>.

Original comment by: lb42

raffazizzi commented 11 years ago

Group: AMBER --> GREEN
Priority: 5 --> 1(low)

Original comment by: lb42

raffazizzi commented 11 years ago

Well, the usual reason for treating as a group of texts what one might otherwise treat as a text divided into divs is that one or more of the included div-like parts has its own front and/or back matter. E.g. a collection of sermons is a text, each sermon is a div; but these are frequently published both separately and as a collection, and in the latter case the title pages, dedications, etc., of each sermon are preserved in the collected edition. What would be a div'd body must necessarily be tagged as a group of texts. But I think the request must relate to some other need, since in my example the superordinate text tag happily bears the @type attribute, with no need to type group, or body for that matter. I have no objection to adding @type to group, however, in the interest of the abstract structure text-group-text-div.

Original comment by: pfschaffner

raffazizzi commented 11 years ago

I see no reason not to do this. I can imagine other, less controversial, contexts in which a single document would contain different groups of texts of different types (e.g. one group of extracts from dramatic texts and one of extracts from poetry).

Original comment by: martindholmes

raffazizzi commented 11 years ago

As with Martin, seems to pass my straw poll test of whether something should get att.type: "Is it repeatable? Yes. Is it reasonable that you might want to classify them in more than one way? Yes."

-James

Original comment by: jamescummings

raffazizzi commented 11 years ago

status: open --> closed-accepted

Original comment by: lb42

raffazizzi commented 11 years ago

Implemented at rev12281

Original comment by: lb42

raffazizzi / TEI-TEST

<group> should be a member of att.typed #128