search

lombardpress / lombardpress-schema

0 stars 2 forks source link

Lemma for additions #116

Open stenskjaer opened 7 years ago

stenskjaer commented 7 years ago

This issue was reported by Majorie Burghart. She wrote:

As I have already told Jeffrey, I beg to disagree with the use of @n to store the desired lemma to be displayed, when the readings are an omission. On principle, using an attribute for that is not fitting well with the idea behind TEI P5: you might need some more precise encoding in this desired lemma, and therefore it should be stored in an element rather than an attribute. Just a basic example, off the top of my head: 

... et sunt IIII<hi rend="superscript">or</hi> 
<app> 
<lem/> 
<rdg wit="#B"></rdg> 
</app> proprietates ...

I'm not particularly recommending this element, but in my own editions I use a with an attribute @type="altLem" (alternative lemma). 

... et sunt IIII<hi rend="superscript">or</hi> 
<app> 
<lem/> 
<rdg wit="#B"></rdg> 
<note type="altLem">IIII<hi rend="superscript">or</hi></note> 
</app> proprietates ...

When this element is present, it is displayed instead of the contents of , which can be useful not only for additions, but also for whatever situation when you want to change the lemma (for instance, when the lemma is veeeery long and you just want to display a couple of words at the beginning and end, with an ellipsis in the middle).

stenskjaer commented 7 years ago

My immediate thoughts on this are:

  1. I think it is a good point that the rendition of the preceding element would not be available in the @n.
  2. Personally I am not convinced by the use of a <note>. A note is not conceptually as close to the lemma as I would like, and adding a @type attribute looks to me a bit like that attribute is used for creating pseudo-elements to fill out gaps in the schema although maybe the element that it is attributed does not really represent what it contains. What I mean is: A <note> is a note, not a lemma. I see how this might lead to "@type-soup".
  3. I don't really buy the argument about the long lemma. As I see it, decisions about how the lemma should be represented are to be made at the processing level, not the encoding level (an example of how I have done this).

Another way of approaching this is to loose the hardcoded text and use a reference in stead. I wonder whether @prev or maybe @corresp could be used. I don't quite know which I prefer right now. @prev has the name going for it, but we use that differently in other contexts (which I think are more correct, namely as a reference to previous element). @corresp on the other hand seems sufficiently vague in the TEI handbook that it could fit ("(corresp) points to elements that correspond to the current element in some way." http://www.tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.global.linking.html). It could then look like this:

... et sunt <seg xml:id="arbitrary-id-asf2o">IIII<hi rend="superscript">or</hi></seg>
<app> 
<lem corresp="#arbitrary-id-asf2o"/> 
<rdg wit="#B"></rdg> 
</app> proprietates ...

The processor would then fetch the content of that element. There may be problems with this approach, but I see the following advantages:

  1. It keeps the annotation on the relevant element.
  2. You avoid typing the same material more than once.