From Lopes' et.al paper on naturalistic programming 1):
DSALs might be a better approach than a general purpose aspect language. The reason is that domain-specific aspect languages can utilise a higher level join point model. We can use a domain-specific language that generates code that is the basis for a higher-level join point model.
We believe that DSALs offer more than just that, an obvious extra bonus is being able to abstract over operational details of the advice, i.e. bringing the advantages of DSLs to aspects. More arguments can be found in the literature, for example in the XAspects paper by Shonle et. al.2)
We have created an inventory of all DSALs we have encountered in the literature. For each language we give a brief description, along with a reference to the corresponding paper.
DSALs have their own workshop series, DSAL, held at GPCE in 2006 and AOSD in 2007, 2008, 2009, 2010, 2011, 2012. Proceedings of various of the workshops, are available in the ACM Digital Library DSAL07, DSAL08, DSAL09, DSAL11, DSAL12