Encoding USAS in TEI

[TomazErjavec]

This issue discusses the non-resloved problems from #202 and #204. The current encoding of USAS in TEI is given in the guidelines, which is arguably ok, even though other possibilites exist (in particular stand-off markup where there are no problems with crossing XML tags but resolving them then gets complicated). Also, it is not yet clear whether retaining per-word USAS tags is sensible in the context of MWEs. These dilemas should be solved here.

The conversion of CoNLL-U with USAS tags into TEI is done by the conllu2tei.pl script. This script is badly written (it first just inserts <name> and <phr> into a temporary TEI and then afterwards tries to resolve conflicts, but does so in a bad way, i.e. it removes <phr> elements even in cases where it shouldn't, in particular phr/name, (arguably) name/phr, and and when a phr is adjecent to name, which is a definite bug. Again, how to make the script better should be discussed here.

[TomazErjavec]

Here is the alternative proposal on how to encode USAS in TEI:

<seg xml:id="ParlaMint-LV_2014-11-04-PT12-264-U1-P1" xml:lang="en" corresp="mt-src:ParlaMint-LV_2014-11-04-PT12-264-U1-P1">
   <s xml:id="ParlaMint-LV_2014-11-04-PT12-264-U1-P1.1" n="1" corresp="mt-src:ParlaMint-LV_2014-11-04-PT12-264-U1-P1.1">
    <w xml:id="tok01" pos="NNP" msd="UPosTag=PROPN|Number=Sing" lemma="Mr.">Mr.</w>
    <w xml:id="tok02" pos="NNP" msd="UPosTag=PROPN|Number=Sing" lemma="President" join="right">President</w>
    <pc xml:id="tok03" pos="Z" msd="UPosTag=PUNCT" join="right">-</pc>
<!-- ... -->
    <spanGrp type="sem">
      <span target="#tok01 #tok02" type="Z1mf,Z3c" ana="sem:Z1"/>
      <span target="#tok03" type="Z9" ana="sem:Z9"/>
    </spanGrp>
  </s>
</seg>

My objections would be that:

• it introduces a completely new construct for linguistic analysis, so far not used in ParlaMint or Parla-CLARIN
• it just postones the problem of conflict with <name> to whatever application or conversion that will try to use both

So, I would advocate sticking to the current encoding but fix the converson script to do a better job.