|
From: | Marco Vassallo |
Subject: | RE: mesh refinement |
Date: | Mon, 24 Jun 2013 08:25:39 +0200 |
> From: address@hidden
> Subject: mesh refinement > Date: Mon, 24 Jun 2013 06:46:43 +0200 > CC: address@hidden > To: address@hidden > > Marco, > > mesh handling functions look very nice, well done! > > From your blog: > > "The code, which is available here, allows us to obtain either a uniformly refined mesh or to specify the list of the vertices we want to be refined." > > I guess you actually meant "cells to be refined", right? > Hi, Yes, sorry, they are obviously cells, as I've shown in the mini tutorial! > "P.s. As you have probably remarked, our functions are not able yet to treat some information available in the e matrix. For example the label of the domain to the left/right of the side elements, but we'll work on it later.." > > The function msh2m_topological_properties can probably do that for you in 2d already, give it a look, extensions to 3D should not be too difficult. > That is a very good news, I was afraid of how we could have done it! > You should now work on a work-plan for the rest of the project. > > I would suggest using TDD: > > 1) First write an example of how you expect the Octave interface to look like, > e.g. rewrite some of the C++/Python tutorials using Octave syntax, to get an idea > of the classes and methods you need to implement > > 2) Define a smaller test for each of the classes/methods you need and define a timeline > for its implementation (be conservative when estimating development times!) > > 3) Start coding! > > c. > > Sounds good to me, but it is not yet clear to me if I have to start with pre-built variational form and to extend it later, or if we directly start with a generic interface for DOLFIN/UFL > P.S. If you can don't miss Carnës talk today about Octave-Forge packaging: https://www.google.com/calendar/embed?src="">> > as he will be talking about how to run configure scripts during package installation. > > > |
[Prev in Thread] | Current Thread | [Next in Thread] |