Re: [Libmesh-users] adaptivity criterion

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

On Sat, 1 May 2010, Boyce Griffith wrote:

> On 5/1/10 1:41 PM, Roy Stogner wrote:
>> 
>> On Sat, 1 May 2010, Boyce Griffith wrote:
>> 
>>> If I do this manually, do I set flags only for active elements, or do
>>> I need to flag all elements?
>> 
>> Not all element, but parents with children to be coarsened need to
>> have the children marked COARSEN and the parents COARSEN_INACTIVE.
>
> Is it OK for some children to be flagged COARSEN and others to be flagged 
> DO_NOTHING or REFINE?  Or do all children need to be flagged COARSEN?

I don't recall for sure, but I think we assert() somewhere that
they're all consistent: flag one for COARSEN and you need to flag the
others and the parent the same way.
---
Roy