On 22 Apr 2014 at 23:33, Vicente J. Botet Escriba wrote:
I'm not concerned with MPL-backward compatibility or the learning curve for users coming from MPL. You have an opportunity to build a new TMP library, the work will be long and hard, but adding more constraints than needed wouldn't help. TMP should be a basic library and other inter-operable libraries must be build on top of it.
Of course, the new TMP should perform better and be easier to learn than MPL. This must be possible as C++14 provides more features and so you will be able to choose the ones that go towards this goal.
What I would like is something that can be proposed to the C++ standard. A little bit more of TMP in the standard would be much better than much more in Boost but outside the C++ standard.
+1 on Vicente's opinion. I see no problem with your current plan Louis, and besides everyone here is a backseat driver and doesn't really have the domain experience that you do. I'd present on the options you've discovered so far at C++ Now as you're already doing, and listen to Joel's opinion, he's very experienced on the topic, and I'm sure he'll make sure you go the right way. Maybe this time next year once you have the TMP14 engine in place and well tested, then would be the time to think about MPL98 compatibility and all that jazz as a second GSoC. Between now and then you'll need to decide on Concepts, sadly Andrew won't be at C++ Now this year as he's the right person to talk to, but I may try to very poorly fill Andrew's place and talk to you myself after your talk. Niall -- Currently unemployed and looking for work in Ireland. Work Portfolio: http://careers.stackoverflow.com/nialldouglas/