[petsc-dev] DMPlex "chart" term

[Patrick Sanan]

On Wed, Oct 11, 2017 at 10:58 AM, Matthew Knepley <knepley at gmail.com> wrote:

> On Wed, Oct 11, 2017 at 4:54 AM, Patrick Sanan <patrick.sanan at gmail.com>
> wrote:
>
>> Another terminology question to help with the docs.
>>
>> What's the origin of the term "chart" in DMPlex? I'm only previously
>> familiar with the term in the context of manifolds (where chart =
>> homeomorphism from an open set in the manifold to an open set in R^n).
>>
>
> That is what it is supposed to be. A chart is coordinates on a patch of
> the manifold. We intend the same thing here in that you can
> locally refer to points with a given name, but they may have a different
> name on another chart (process). The names are matched up using the PetscSF.
>

So do you mean that you think of the DMPlex (thought of just as a set of
points) as some kind of a "manifold over the integers" and consider an
"atlas" with one patch per proc, with the SF defining the "transition maps"?


>
>> In terms of the use in DMPlex, is there anything wrong with thinking of
>> "chart" as a shorter way to say "interval of integers, closed on the left
>> and open on the right, e.g.  [pStart,pEnd)"?
>>
>
> No, since we depend on continuity of point names almost everywhere.
>
>   Matt
>
> --
> What most experimenters take for granted before they begin their
> experiments is infinitely more interesting than any results to which their
> experiments lead.
> -- Norbert Wiener
>
> https://www.cse.buffalo.edu/~knepley/ <http://www.caam.rice.edu/~mk51/>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20171011/0f6a10e6/attachment.html>