Re: [Libmesh-devel] Random seed for MeshTools:Modification::distort

[Roy S. <roy...@ic...>]

On Mon, 9 Jan 2012, Vetter Roman wrote:

> May I ask if something of the proposed kind will be implemented
> anytime soon? If not, I will have to continue using my own hacked
> "branch" of libmesh instead of the trunk... ;-)

Probably not soon.  I hate the idea of leaving you stuck on a
permanently forked branch, though...

Perhaps we could avoid using a sequential PRNG in distort()
altogether?  Replace the std::rand() subroutine calls with some kind
of function calls that lack internal state, just hashing the initial
seed with the previous node coordinates somehow?
---
Roy