Robert Marion wrote:

...

i have written some primitives to be used with Boost.Accumulators, and i'm considering to make them available. these files are mostly *.hpp and depend on boost headers. under what license is it preferable that I distribute them under?

Hi, That depends on your intention. The BSD style and MIT style licenses are permissive, attribution licenses. The Boost is also a permissive license. GPL and LGPL are constrained so that anyone using your code will have to make theirs open source (depending on how it is used, in the later case).

If you are writing *for* Boost, the responsible thing to do is use the Boost license. If not, and you are looking for a permissive attribution license, then I suggest either the BSD or MIT style because they are commonly known. There are many other licenses, such as Apache, CDDL, etc. If you want to guarantee your code remains open source, then GPL or LGPL are a good choice (but many persons/companies will choose not to use your code if it falls under one of those licenses).

Yours, Robert Marion

http://www.palamida.com

Hi, Thanks for the advice. Am I writing *for* Boost? The project I have are tools for estimation of dependent data. I used the code from Boost.Accumulators (and Boost.Accumulators_filters) as templates with modifications that meet my needs, so it's nothing crazy, but still probably short of the requirements for submission to Boost. Here's a snippet: typedef accumulator_set< double, stats< tag::acvf<>,//uses tag::delay<> from accumulators_filters library tag::acf<>, tag::integrated_acvf<>,//a.k.a variance of an autocorrelated process tag::percentage_effective_sample_size<>, tag::standard_error_autocorrelated<>, tag::standard_error_iid<>

acc_type;

From the submission guidelines, it appears that there's 1) a preliminary submission to the vault, and later 2) a formal review. Not sure what submission to the vault exactly entails... Does it sound like I should

put it in the vault, and refine the library over time? Thanks.