- 积分
- 44
- 注册时间
- 2005-3-18
- 仿真币
-
- 最后登录
- 1970-1-1
|
发表于 2006-8-26 22:59:09
|
显示全部楼层
来自 日本
Thanks for your reply
As you suggested, maybe I will try to find the book 《有限元法与面向对象编程》.
I have compiled successfully getFem++, libmesh, Tahoe in WINDOWS. If in linux, no prblem of course.
The character of getFem++ is its generalized element library as well as use of strategy template. It can be seen from its source that the developer have a deep understanding of finite element analysis. But I don't like its style, i.e., as to myself, readability of the source code is not good.
Doumentation and readability of dealII is extremly good. But their element libaray is not extendably! Oh! what are those guys doing?
The idea of libmesh that seperate the management of mesh operation from FEM calculation is quite advanced.
About Tahoe, its style is good. It even includes somwthing like mesh free method, crystal plasticity. Quite attractive then?
Elmer is not object-oriented soft.
My job now is programming FE soft. I look forward to your review about OOFE. Anyway, program in object-oriented soft is generelly hard to understand than that of FORTRAN.
P.S.: Seems not like template. Try to read "Modern C++ Design: Generic Programming and Design Patterns Applied" of Andrei Alexandre. |
|