On 18 August 2014 16:32, Robert Ramey
I'm curious about this metadata idea.
Maybe there has been a discussion about it and I didn't see it.
It's been discussed on several occasions, often in passing. When I first started adding the metadata to libraries, I posted about it: http://lists.boost.org/Archives/boost/2014/02/211799.php
But in any case I have a couple of questions:
a) what is the immediate intended purpose and benefit?
The immediate benefit is that it lets people update their library details without modifying the website.
b) what are some possible future intended purposes and benefits?
Creates a single location for the data, which is currently duplicated in several different places, and is often inconsistent. We might also add more data, such as expected failures.
c) why was json chosen rather than some xml schema. The reason I ask this is that your documentation build system is based on xml transforms.
My documentation build system? If it was up to me, the documentation build wouldn't have been based on xsl.
It seems to me that (not having answers to the a and b above) that a more natural choice for something like would be some xml schema.
It was originally in xml, because that's what the website uses. But someone asked for json, and that seemed fine to me as I think people find it easier to read and write.