Reece Dunn wrote:
I agree that Boost.Spirit, Boost.Serialization and my library overlap to certain degrees.
IOW, CSS would be unnecessary complication for Output Formatter at this stage. I like it this way:simple to understand and use. Formatting options are reasonably limited.
This is the direction I am currently taking. Possibly, at some point in the
future, this library could become the output complement to Spirit (allowing
advanced code generation), as well as Spirit simplifying the binding of data to the parser. But then again, I don't know how feasible that would be.
Personally, I don't see much overlap at all between, serialization, io streams, spirit and output formatting. I see them a more complementary. Of course serialization uses streams so might work with io streams library. Serialization uses spirit to implement xml archives and it is possible someone might want to make an "annotated" archive and so want to exploit output formatting facilities. But in general these are orthogonal functions that can be composed at will. So I would be careful about making the output formatter library bigger. Bigger isn't necessarily better. I think its better if it focuses on doing one thing better than any other library does it and presents a minimal interface so that can be easily composed with related but different concepts. Robert Ramey