It looks like what this lecture is leading towards is a declarative language, where you declare what operations you want to perform and the back-end figures out the most efficient way of doing so. The canonical example (at least to my knowledge) is SQL. Users describe how they want their database to look (the schema) and create queries which are really just a description of how to pull data out of this database. Implementations are free to represent the back-end, the tables, however they would like and this has led to a variety of implementations each with their own strengths and weaknesses.
This comment was marked helpful 1 times.
asinha
There are two main ways to implement a Domain-Specific Programming System. The first way is to develop a domain specific language as a standalone language, which in this case would involve a process close to what @tchitten commented on above, and SQL and Matlab are examples of this process. The second method involves embedding the domain specific language in a general programming language, and examples of this process are the C++ library and Lizst which is described in the slides ahead. This second method usually involves facilitation where the compiler can be modified so that the parsing part of the compiling process is utilized, but the optimization and code generation phase is modified.
It looks like what this lecture is leading towards is a declarative language, where you declare what operations you want to perform and the back-end figures out the most efficient way of doing so. The canonical example (at least to my knowledge) is SQL. Users describe how they want their database to look (the schema) and create queries which are really just a description of how to pull data out of this database. Implementations are free to represent the back-end, the tables, however they would like and this has led to a variety of implementations each with their own strengths and weaknesses.
This comment was marked helpful 1 times.
There are two main ways to implement a Domain-Specific Programming System. The first way is to develop a domain specific language as a standalone language, which in this case would involve a process close to what @tchitten commented on above, and SQL and Matlab are examples of this process. The second method involves embedding the domain specific language in a general programming language, and examples of this process are the C++ library and Lizst which is described in the slides ahead. This second method usually involves facilitation where the compiler can be modified so that the parsing part of the compiling process is utilized, but the optimization and code generation phase is modified.
This comment was marked helpful 0 times.