SEARCH
0-9 A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
Prev | Current Page 556 | Next

Steven Kelly and Juha-Pekka Tolvanen

"Domain-Specific Modeling"

Before you embark on that route and rediscover the only workable
solution??”having some intermediate form from which you can generate both
languages??”stop for a second and remember that you already have precisely such an
intermediate form: the models.
Automatic Translation and Intermediate Formats There have been
attempts in some DSM tools to force the use of such an intermediate form between
the models and the code. The background to such attempts is however not from the
world of DSM, but rather from UML. In UML, the domain is object-oriented
languages, and there is no expectation that working applications will be generated
from models. Instead, it is suf?¬?cient for the UML tool to generate the class and
function headers. As there is no problem domain semantics in the models, and hence
no knowledge of what the code should actually do, the generation simply maps a
graphical description of object-oriented classes into a textual format. At that level,
object-oriented languages are similar enough that simple mappings can be made, and
where the mapping is less than ideal, it does not matter anyway, since nobody is going
to execute the code.
Coming from such a background, where all modeling languages are UML and
all programming languages are object-oriented, it is tempting to envisage DSM
GENERATOR STRUCTURE 303
code generation as being ?¬?rst a transformation into UML, then a transformation per
language from UML into code.


Pages:
544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568
hotel jelenia góra Russian bride Free English grammar and study guid powiekszenia wielkoformatowe counter strike 1.6