On 6/17/2013 10:29 AM, Andrey Semashev wrote:

I know, C++11 has many fancy features and all, and I'm all for its adoption too. But Boost also serves practical purpose, and if people can't use your library then that just limits its usefulness. So unless you trying to make some academic work here, the library should be more portable.

Seems an undue constraint on a new library to me. Adding a new C++11 library harms no one that doesn't have have a C++11 compiler. If you need a C++03 version of a C++11 only library and you feel it is not an unreasonable amount of work to provide you could just fork the library into your own github and do the backporting.

On 6/17/2013 10:55 AM, Hartmut Kaiser wrote:

...

...

I know, C++11 has many fancy features and all, and I'm all for its adoption too. But Boost also serves practical purpose, and if people can't use your library then that just limits its usefulness. So unless you trying to make some academic work here, the library should be more portable.

Seems an undue constraint on a new library to me.

Adding a new C++11 library harms no one that doesn't have have a C++11 compiler.

If you need a C++03 version of a C++11 only library and you feel it is not an unreasonable amount of work to provide you could just fork the library into your own github and do the backporting.

Whether this is good enough for Boost will be decided during the review and not by individual interests.

Makes sense.

On Monday 17 June 2013 10:39:10 Michael Marcin wrote:

On 6/17/2013 10:29 AM, Andrey Semashev wrote:

...

I know, C++11 has many fancy features and all, and I'm all for its adoption too. But Boost also serves practical purpose, and if people can't use your library then that just limits its usefulness. So unless you trying to make some academic work here, the library should be more portable.

Seems an undue constraint on a new library to me.

Adding a new C++11 library harms no one that doesn't have have a C++11 compiler.

If you need a C++03 version of a C++11 only library and you feel it is not an unreasonable amount of work to provide you could just fork the library into your own github and do the backporting.

I wasn't suggesting making the library strictly C++03-compatible. My main point was that the library has to be compatible with _todays_ and even better - _yesterdays_ compilers to be actually useful. My choice of "reasonable" time frame is 3-5 years, as I mentioned. That includes VS 2008 and VS 2010, which already had some C++11 features, including rvalue references. Making VS 2013 absolute minimum is a no-go, IMHO.

Daniel James wrote:

On 17 June 2013 18:17, Andrey Semashev wrote:

...

On Monday 17 June 2013 10:39:10 Michael Marcin wrote:

...

If you need a C++03 version of a C++11 only library and you feel it is not an unreasonable amount of work to provide you could just fork the library into your own github and do the backporting.

I wasn't suggesting making the library strictly C++03-compatible. My main point was that the library has to be compatible with _todays_ and even better - _yesterdays_ compilers to be actually useful. My choice of "reasonable" time frame is 3-5 years, as I mentioned. That includes VS 2008 and VS 2010, which already had some C++11 features, including rvalue references. Making VS 2013 absolute minimum is a no-go, IMHO.

The portability requirements disagree.

http://www.boost.org/development/requirements.html#Portability

+1 - agree. This has always been the standard and it has worked well. In a practical sense, if one makes new a C++11 only package - no current user is impacted. If someone want's to use it, they'll just have to find a way to upgrade. This is not an unfair, nor unrealistic policy. And it's better for C++. If some submission is very populer but only supports C++11, it will help drive C++ forward - which to me is one of the principle goals and accomplishments of Boost so far. And the work in making a new submission backward compatible is basically wasted and the value of this work diminishes to nothing over time. Let's keep Boost looking forward ! Robert Ramey

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

On Jun 17, 2013, at 8:47 AM, Jonathan Wakely wrote:

On 17 June 2013 16:29, Andrey Semashev wrote:

...

I know, C++11 has many fancy features and all, and I'm all for its adoption too. But Boost also serves practical purpose, and if people can't use your library then that just limits its usefulness. So unless you trying to make some academic work here, the library should be more portable.

And if the maintainers have to spend twice as long implementing it to be C++03-compatible and it isn't ready to be included in Boost until next year that also limits its usefulness, to *everyone* not just the C++03 crowd.

I'd like to widen this discussion; it's shouldn't be "C++03" vs. "C++11" . The correct question is "How should boost support/deal with the changes in C++?" For a long time, that was C++03 (and TR1). Now we have C++11. A year from now, (seriously - next June) we will have C++14 - and some compilers have (partial) C++14 support today. What is the best way forward for the people who use and/or develop boost libraries? -- Marshall Marshall Clow Idio Software mailto:mclow.lists@gmail.com A.D. 1517: Martin Luther nails his 95 Theses to the church door and is promptly moderated down to (-1, Flamebait). -- Yu Suzuki