Schema Evolution publication categorizer

This paper presents ATL (ATLAS Transformation Language): a
hybrid model transformation language that allows both declarative and
imperative constructs to be used in transformation definitions. The paper
describes the language syntax by using examples. Language semantics is
described in pseudo-code and various optimizations of transformation
executions are discussed. ATL is supported by a set of development tools such
as an editor, a compiler, a virtual machine, and a debugger. A case study shows
the applicability of the language constructs. Alternative ways for implementing

Interoperability of heterogeneous data sources has been extensively
studied in data integration applications. However, the increasing number of
tools that produce data with very different formats, such as bug tracking,
version control, etc., produces many different kinds of semantic heterogeneities.
These semantic heterogeneities can be expressed as mappings between the tools
metadata which describe the data manipulated by the tools. However, the
semantics of complex mappings (n:1, 1:m and n:m relationships) is hard to
support. These mappings are usually directly coded in executable

We consider in this paper that a DSL (Domain Specific Language)
may be defined by a set of models. A typical DSL is the ATLAS
Transformation Language (ATL). An ATL program transforms a source
model (conforming to a source metamodel) into a target model (conforming
to a target metamodel). Being itself a model, the transformation
program conforms to the ATL metamodel. The notion of metamodel is
thus used to define the source DSL, the target DSL and the transformation
DSL itself. As a consequence we can see that agility to define

Mapping between heterogeneous
data is a central problem in many dataintensive
applications. In particular, using
one mapping language causes serious
limitations and makes mapping management
difficult. In this paper, we propose a solution
that can better control the trade-off between
genericity, expressiveness and efficiency of
mappings. Our solution considers mappings
as models and exploits specific mapping
engines. We define model weaving as a
generic way to establish element
correspondences. Weaving models may then
be used by a model transformation language