Bumping BGL's minimum C++ standard: input wanted on target version

[Becheler]

Hi Graphlings! 🚀

During our modernization workshop in Paris (see #466), one of the most frequent discussion points was the C++ standard version BGL should require. Modernization is intuitively at odds with C++14 😉

Some context worth knowing:

• Boost.Graph is already C++14
• A lot of API modernization can be done without bumping the minimum.
• Some modernization changes would land cleaner on a newer standard
• We don't yet have a clear enough picture of our users' constraint space to inform our discussion about a potential baseline bump. For example, it appeared that automotive industry users will be locked in C++17 for a while with MISRA.

Your answers will inform our discussion about a potential baseline bump.

We'd deeply appreciate it if you could take a moment to answer this simple poll. Thanks!

[nigels-com]

I voted C++20. For our new project at work we're C++23. Assuming that BGL1 remains an option for a good while, take the opportunity to come forward. 😀 I'm only a potential user of BGL.

[Becheler]

Thank you @nigels-com ! 😄
We really appreciate you took the time to answer !

[rcasellas]

Hi

Thank you for the poll. I have been actively using boost graph since 2002 but , fortunately, I am free to choose the most recent c++ compiler to use as I see fit. I voted for C++20

[Becheler]

Thank you @rcasellas ! I am happy for you to learn you are not stuck in C++11 😄 Or C++17 like apparently many people dealing with MISRA 😅
Out of curiosity, in what sector do you work for having used BGL since 2002 ? :)

[rcasellas]

sure :) I started using boost graph for my Ph.D. thesis (ENST Paris but not living in France anymore) in the context of network control and management and traffic engineering. I am still using it in a project that has evolved from an IETF Path Computation Element (PCE) to an SDN controller for optical networks. both boost graph and boost asio have been key for these developments :)

[Becheler]

Wahoo beautiful arc ! We may have been neighbors at some point, I was doing my PhD at Gif-sur-Yvette/Laboratorie de Mathematiques d'Orsay, and my PI was from Polytechnique 😉
If you ever hit rough edges, missing algorithms, or API frustrations along the way, please open issues or ping me here. Feedback from someone running BGL in production for 23 years is gold for us 🙏
Also have you seen the new documentation at https://becheler.github.io/graph/graph/index.html ? Your feedback is welcome on #478 ❤️