Hey,

> Whether we can use boost.proto framework for designing a EDSL for
> distributed component placement Project or we have to implement our own
> expression templates and design our own evaluation methods?

Are you talking about this project idea: 

https://github.com/STEllAR-GROUP/hpx/wiki/GSoC-2017-Project-Ideas#distributed-component-placement

?

Assuming that's the case, I think that using Boost.Proto would be a marvelous 
overkill for this. 

One of the main tasks for this project would be to design a EDSL in the first 
place, i.e. defining a syntax which would encapsulate all possible use cases. 
This certainly has to be done with having the C++ syntax in mind, as for a EDSL 
we can't escape the language. Once this has been done we can chose the way to 
implement the EDSL, where this decision will very much depend on the complexity 
of the language designed. I would expect the EDSL to be not too complex, after 
all its purpose would be to describe a set of fairly narrowly constrained 
placement and distribution options. 

But we might decide that the EDSL is too complex for a straightforward 
implementation, in which case we might chose an adequate library for its 
implementation. 

Overall, I wouldn't suggest to nail this particular implementation detail at 
this stage of the project.

HTH
Regards Hartmut
---------------
http://boost-spirit.com
http://stellar.cct.lsu.edu




_______________________________________________
hpx-users mailing list
[email protected]
https://mail.cct.lsu.edu/mailman/listinfo/hpx-users

Reply via email to