Pen-Based Interaction for Intuitive Music Composition and Editing - Intelligent Music Information Systems: Tools and Methodologies - page 270

Figure 5. Architecture of the generic system

The graphical and pen-based editing functions
and the hand-drawn stroke analyzer are domain-
independent components ; they do not need to be
rewritten for each pen-based system. The develop-
ment of such systems with our approach only re-
quires realizing the domain-specific components :
writing the interpretation rules (embedded under
a compiled form), designing the necessary hand-
drawn shape recognizers (which can be reused
from one system to another) and specifying the
graphical information. The components of our
system are well separated and can be modified
and adapted independently of the others.
We present these components more in detail
in the next subsections.

• The modeling of the composition conven-

tions: As the interpretation process is eager,
the system has to deal with incomplete
documents; a way to do so is to model how
a document is typically drawn, for instance,
which element after which other.

• The representation of the document

spatial structure: It is necessary to exploit
the relative positioning of the document ele-
ments in order to model, on the one hand,
in which document structural contexts an
element can be identified, and, on the other
hand, which of these contexts are generated
due to the creation of an element.

• The driving of the recognition process

by the document context analysis: We
have stated that it is not possible to have a
unique recognizer for all the symbols that
a document can contain. By exploiting the
document context of an element, the system
must reduce the likely symbols and choose
which dedicated recognizers to use.

a formalism for eager Interpretation
of structured documents

The main component of our approach is a for-
malism which aims at defining how to eagerly
interpret the elements of structured documents
from various domains. The goal is to propose a
formalism which is as generic as possible in order
to be able to deal with a large panel of documents.
We previously (Macé et al., 2005a) defined the
four basic concepts that are, in our opinion, as-
sociated to the modelling of eager interpretation
of online structured documents:

• The pen-based human-computer inter-

action: As the result of the interpretation
process is directly displayed on the screen,
the user is aware of it. Consequently we can
take into account the human-computer inter-
action and exploit it on an as user-friendly
way as possible.