Rather, it's about an approach to library design that errs on the side of providing the minimum possible, rather than the maximum potential permutations of functionality.
Of course, age, learning, evolving paradigms, and backwards-compatibility all tend to result in accretion of functionality. And there's nothing going to happen in 1.9.x to start removing code. But as of 1.10 I'd like to rewrite several popular libraries to address the crustification.
To guide me in this, I'm looking for users to write in and make their complaints/requests.
Obvious (to me) candidates for face-lifts include:
- FileSystem Library
- System Library
- Windows Registry Library (see include/winstl/registry/*)
- Windows Controls
- Error Library (in fact this is already happening)
- Containers Library
- String Library (see this post)
Do you have any more? If so, please let me know, either here or on the STLSoft newsgroup.
No comments:
Post a Comment