TEI Syntax and Metalanguage Committee
                   Statement of Work
               Document Number: TEI MLW11
                      February 1990
 
This Committee has responsibilities in two areas.
 
The first is the establishment of guidelines for the other committees
with respect to the formalism to be used in generating and specifying
the encoding scheme that will be the result of the Text Encoding
Initiative. The default assumption as the Committee begins its task is
that the formalism to be used will be the Standard Generalized Markup
Language (SGML). The expected results of this part of the work are:
 
 o  an endorsement of SGML as the basis for the TEI,
 o  an indication of how SGML is to be used and what features are to be
avoided, and
 o  an indication of whatever extensions or conventions are necessary
to encode aspects of the TEI that are not explicitly handled by SGML;
it is hoped that these will be few and minor.
 
 
The second area of responsibility is the development of examples of
translations between the scheme developed by the TEI and existing
``important'' text encoding schemes. Schemes to be dealt with will be
chosen on the basis of their importance to language industries or to
research and scholarship (as indicated by breadth of use) and the
diversity of the techniques needed to accomplish the translation. The
expected results of this part of the work are:
 
 o  a categorization of existing encoding schemes, with an indication of
those representative schemes to be used as examples,
 o  an identification of a descriptive technique (or descriptive
techniques) to be used in specifying the translations,
 o  for each of the examples, a description of the translation from the
TEI scheme to the target scheme, and a description of the translation
from the target scheme to the TEI scheme, and
 o  for at least some of the most important examples, a working
implementation of the translations.
 
The mappings between the TEI scheme and other schemes should have the
following properties.
 
 o  Mappings from some other scheme to the TEI scheme should be
accomplished without loss of information from the encoding, for all
features in the source for which a corresponding feature exists in the
TEI scheme.
 
 o  Mappings from the TEI scheme to other schemes are expected, in
general, to lose information because we expect the TEI scheme to be a
rich one that includes features not present in the target scheme. These
mappings will be restrictions or filterings.