Here's a counterpoint for you: Boost is maintained by some of the leading experts in C++ in the world.
Over the years I've received a lot of advice from Boost authors. Some of it has been directed to me and some of it has come from reading the Boost mailing lists and applying something. Originally the only way to use the CMW was via this web interface. (That interface is no longer active/live.) Someone on a Boost list encouraged me to develop an interface that could be easily integrated into build environments. The CMW is a new Boost Serialization library. It has support for some Boost libraries that the serialization library that's in Boost doesn't have. Originally I had very little open-source code... today I have over 5,000 lines of it and growing. I've gotten a lot of helpful input from many highly regarded C++ people: Bjarne Stroustrup, James Kanze, Boost authors, people here and on other forums. And Telastyn noted that the CMW is also free.
Your project is maintained solely by you, and having existed for over 9 years, still has glaring code quality issues by even your own admission.
What do you think people are liable to trust more?
[/quote]
I'm one of the maintainers of the CMW, but not the only one. I don't argue that the CMW has no room for improvement. That's why I started the thread. The point is the serialization library that's in Boost is losing momentum. I wouldn't be working on this if I didn't believe there was a changing of the guard taking place. The older approach worked OK for years, but now there's a better (more portable and sustainable) way to do it. We're out here proving it to people. The remarks about how the serialization library in Boost hasn't been well maintained is getting at how sustainable that model is. A business model matters if you want a helpful partner that will be here for as long as your project lasts. Is Google going to release a version of Google Docs that you install locally? Is Microsoft going to stop developing Office 365?