#381 French example of <metDecl> seems wrong

AMBER
closed-fixed
Martin Holmes
5
2012-06-12
2012-03-09
Martin Holmes
No

This example is given for <metDecl>:

<metDecl xml:id="fr_ip" type="met" pattern="((AATAAT|AATAATA\)">
<metSym value="T">syllabe tonique</metSym>
<metSym value="A">syllabe atone</metSym>
<metSym value="|">pause métrique</metSym>
</metDecl>

The pattern has two opening parentheses, and only one closing parenthesis. It also ends with a backslash, which is not defined as a <metSym>/@value; it it's intended to escape the following parenthesis, then we're left with two unmatched opening parentheses. One way or another, I think something is wrong with this.

I think the pipe character, which is defined as pause métrique, should be escaped with a backslash too. So I think the intended pattern should be:

<metDecl xml:id="fr_ip" type="met" pattern="AATAAT\|AATAATA">
<metSym value="T">syllabe tonique</metSym>
<metSym value="A">syllabe atone</metSym>
<metSym value="|">pause métrique</metSym>
</metDecl>

I'm assigning this to myself, but I'd like someone else to confirm I'm not missing something obvious here -- someone good at regexes (Stuart? Sebastian), and if possible someone whose knowledge of French poetry would enable them to confirm the likely intention of the original.

Discussion

1 2 > >> (Page 1 of 2)
  • Lou Burnard
    Lou Burnard
    2012-03-09

    I think Martin's correction is correct; however I've copied the query on to the tei-fr list to see if anyone there has a better idea.

     
  • Martin Holmes
    Martin Holmes
    2012-03-09

    Actually I think there's a broader misunderstanding at work in many of the examples of <metDecl>, and a weakness in the whole method of encoding. The idea, as I understand it, is that the <metDecl> pattern is meant to supply broad constraints with which metrical structures in the text conform. So you would have, for instance:

    - in the header, three <metDecl> elements which define possible patterns to which poetic lines are expected to conform

    - in the text, poetic lines, each of which has an @met attribute which shows an actual pattern which conforms with one of the <metDecl> elements in the header.

    However, as far as I can see, there's no way to say, in the context of the line, which <metDecl> pattern a given line is supposed to conform to (at least, no such method is documented). So if I have, for instance:

    <metDecl xml:id="type1" pattern="[abc]+">...</metDecl>

    <metDecl xml:id="type2" pattern="a[bc]+>...</metDecl>

    and then I have in the text, a line like this:

    <l met="acbc">...</l>

    there's no documented way for me to say that this line is an instance of #type1 or #type2. In this manufactured example, it conforms to both.

    I believe there should be a documented method of linking a block element with metrical structure to the @xml:id of a <metDecl> element. This may consist of @corresp, but even so it should be documented; actually I think it should consist of a new @metDecl attribute.

    Am I misunderstanding or missing something here? I haven't used these metrical structures much in P5.

     
  • Lou Burnard
    Lou Burnard
    2012-03-13

    • milestone: --> AMBER
     
  • Lou Burnard
    Lou Burnard
    2012-03-13

    I interpret the use of metDecl rather differently. It's not meant to define exhaustively all the metrical patterns you may find, it's just defining the syntax you use to express them. However, the French example is still wrong!

     
  • Lou Burnard
    Lou Burnard
    2012-03-26

    An Expert on French Metrics writes:

    Je ne connais pas la syntaxe de metDecl, mais il me semble que l'exemple désigne l'alexandrin (6+6) avec une syllabe atone à la fin. Donc la pause métrique au milieu va bien. La seule chose qui me frappe c'est qu'il y a quatre "T" dans chaque vers. Dans une ligne individuelle, on peut bien trouver ce schéma; par contre, le schéma général de l'alexandrin n'a que
    deux toniques fixés: avant la pause métrique et à la fin de la ligne. Donc je m'attendais plutôt à:

    AAAAAT|AAAAT(A)

    Le (A) à la fin est optionnelle.

     
  • Martin Holmes
    Martin Holmes
    2012-03-26

    I've fixed the French example in rev 10187, following the expert testimony reported by Lou, but I'm leaving this bug open because I'm still not sure that the usage of @pattern is clear. Also, since it's a regex, we should presumably define which regex flavour it's supposed to be.

     
  • James Cummings
    James Cummings
    2012-04-15

    I think Lou's point is that metDecl defines the overall syntax one is using, not a specific pattern. (i.e. whether you are using As and Bs or +s and -s.

    Where it becomes a problem is if you are using As and Bs in one place and +s and -s in another place? Then Martin is right that we haven't documented a good way of saying 'use that metDecl' not this one. Would not an acceptable answer be to use the @decls attribute on the surrounding <div> or whatever? That can point to the @xml:id of the appropriate metDecl. Mixing different formats of @met inside a single poem/section/whatever surely must be considered poor practice, no?

    I agree that the example is wrong, but that has been fixed.

    -James

     
  • Martin Holmes
    Martin Holmes
    2012-06-12

    • status: open --> closed-fixed
     
  • Martin Holmes
    Martin Holmes
    2012-06-12

    Added an example and explanation to section 6.5 Metrical Notation Declaration showing how to declare and use two <metDecl> elements. Committed in rev 10478. Closing the ticket (with fingers crossed that my commit doesn't break anything).

     
  • Martin Holmes
    Martin Holmes
    2012-06-12

    Correction: 10479 (typo in my XML broke the build).

     
1 2 > >> (Page 1 of 2)