Salut Thibaut,
On 2019-04-29 5:38 p.m., Thibaut Cuvelier wrote:
Hi Stefan,
The two tag sets serve a similar purpose, but are very different in
their implementation: the Boost-based set looks like an equivalent to
refentry & co. (the base item is an API/library), while the one I
proposed on GitHub is rather based on the synopses (the base item is
either a class, an enumeration, a macro, or a namespace — the closest
item to Boost root).
Also, the Boost extension adds many many tags (77) ithout being
similar to what already exists in DocBook, while mine tries to use
fewer tags (18, 15 with a generic synopsisinfo) and to mimic the
existing semantics.
The final difference I see is that the Boost extension is really
revolving around C++ (with concepts like headers, which are not
present in that many languages), while I tried to be as
language-agnostic as possible while being able to model as many things
as possible.
Just to highlight a few differences between both :). Actually, I had a
look at BoostBook before choosing to design something else.
I'm not going to argue on technical grounds for either side. However, I
think it would be a big mistake to underestimate the importance of a
community that is using a given technology, and thus is willing to help
maintain it.
BoostBuild is fully implemented and is part of the documentation
toolchain used by many Boost components. Thus, when I mentored the
project to migrate to DocBook 5 and eventually merge it into DocBook
"mainstream", my hope was to enable the wider DocBook community to use
it, making the tool and the project easier to maintain and evolve. This
isn't a project to be taken on by a single person.
Cordially,
Stefan
--
...ich hab' noch einen Koffer in Berlin...
