home *** CD-ROM | disk | FTP | other *** search
- Path: newsfeeds.ans.net!btco!newsadm
- From: Shalom Reich <sqr1874@acf4.nyu.edu>
- Newsgroups: comp.lang.c++
- Subject: Re: Implementing A New Project
- Date: Mon, 04 Mar 1996 12:56:29 -0500
- Organization: Bankers Trust Company
- Message-ID: <313B2ECD.6745@acf4.nyu.edu>
- References: <4herul$mqf@harbinger.cc.monash.edu.au>
- NNTP-Posting-Host: algsvw0058.btco.com
- Mime-Version: 1.0
- Content-Type: text/plain; charset=us-ascii
- Content-Transfer-Encoding: 7bit
- X-Mailer: Mozilla 2.0GoldB1 (WinNT; I)
-
- Andrew Lowe wrote:
- >
- > [snip]
- >
- > The way the program works is that there are
- > points in space called nodes. Nodes usually have an number and a set of
- > coordinates. These are not physical, they serve to indicate a start or end
- > point. In turn members, such as beams and columns, are defined as running
- > between points. Members have properties such as cross sectional area, moments
- > of intertia and the like. They are also made of a material such as steel. The
- > members also have loads applied to them.
- >
- > [snip]
- >
-
- I had some difficulty thinking about nodes. They don't seem to represent real
- life *things* but appear to be artifacts of the implementation. Not knowing
- anything about structural analysis, I would think that "nodes" are attributes
- of members. Is it even necessary for a member to "know" where it is in 3D
- space or is it sufficient to know what other members it is attached to and the
- points in "itself" to which those members are attached? IOW, it appears that
- your current design takes a global, "stucture oriented" view of the world.
- Would it be possible to convert that to a "member oriented" view of the world?
-
- Am I making myself clear?
-
- E-mail me if you think this is helpful and want to follow up.
-
- S. Reich
-