<diehlpk_work>
hkaiser, jbjnr For the JOSS paper we got following remark
<diehlpk_work>
The paper does not have a clear statement of need -- what problems is this software designed to solve (note that the paper lists the benefits of using the software)
<diehlpk_work>
hkaiser, jbjnr Can one of you take care of this?
<diehlpk_work>
ms[m], I updated the namespaces in the latest version of the paper
<diehlpk_work>
Once we addressed the issue above, I will push it to JOSS
<ms[m]>
I'm not 100% sure it's a good idea though... there's overlap between the hpx execution policy and the kokkos execution policy (they both have an execution space instance)
<ms[m]>
I now use the one from the hpx execution policy, but that currently loses some information in reconstructing a new kokkos execution policy using the execution space from the hpx execution policy...
<ms[m]>
in any case, it's doable but may lead to confusing situation, may be fine for simple cases though
<hkaiser>
ms[m]: COOL!
<ms[m]>
OF COURSE ;)
<hkaiser>
we have to find our way here... so this is a good first step
<ms[m]>
keep in mind that there's also `hpx::kokkos::parallel_for_async` which has 100% the same signature as the plain `Kokkos::parallel_for` except it returns a future
<ms[m]>
not as impressive for claiming compatibility with the standard etc etc but it fully preserves the meaning of kokkos execution policies
<ms[m]>
hkaiser: btw, you haven't tried this out on windows, have you?
<hkaiser>
ms[m]: I have not
<hkaiser>
the kokkos build system does not support windows
<ms[m]>
hkaiser: nod, just curious (I know they started adding windows support, but I suppose it's not ready yet)
<hkaiser>
ms[m]: I have not tried recently, though