Demeter Seminar 

Monday 10am - 12 noon
149 Cullinane Hall

April 7, 2003

Topic: XAspects by M. Shonle and A. Shah 
       Presentation by Karl Lieberherr

http://www.ccs.neu.edu/research/demeter/xaspects/

The vision behind XAspects is that each major application domain
needs both a set of cooperating domain-specific languages and a set of
cooperating aspect-specific languages to optimally express the
programs in that domain. XAspects provides infrastructure to
define multiple class dictionaries that share classes to define
multiple views using domain-specific languages and, more
importantly, it defines infrastructure to define new aspect-specific
languages.

It could be argued that it is not feasible to introduce many
little aspect languages that the programmers have to work with. But the
inherent complexity of a project probably requires that the programmers
address the issues of composition of aspect languages anyway so why
not do that through a conscious language design that addresses the
aspect composition issues systematically. Without such an approach,
the programmers will have to do this at the lower level of a general
purpose aspect language.

XAspects translates all class dictionaries and aspect-specific languages
to AspectJ and uses AspectJ as the weaving engine. Currently,
XAspects builds heavily on DAJ
http://daj.sourceforge.net/
and supports the Class Dictionary aspect (.cd) and the Traversal aspect
(.trv)
for programming traversal-related concerns. The synchronization
aspect (COOL) has also been added in projects and a design for marshalling
and remote invocation (RIDL) has been done.

The purpose of this meeting is to discuss the advantages and disadvantages
of XAspects and to identify topics that need to be addressed to make
XAspects more useful.

-- Karl
PS. I like this work because it is a further development of DemeterJ
where we have a small number of aspect languages, including COOL and RIDL.