Nancy,
Thanks for answering this question, I was wondering the same thing myself.
The bummer is that this means that the code for the divergence and curl
assume a cartesian coordinate system. Is it not possible for the
"connections" information to be used to include the metric terms for
calculating the divergence and curl in non-cartesian coordinate systems?
Any idea how much work it would be?
Tom
On Tue, 20 Jun 2000, nancy collins wrote:
>hi,
>
>i've worked with dx and lots of data in different kinds of
>coordinate systems. you do eventually have to convert
>it to x,y before you actually plot up the results, but you
>can choose how much of your computation you want
>to do in cylindrical coordinates before you convert.
>you don't have to do anything special to work in cylindrical
>coordinates - all the modules will work just fine as is.
>
>when do you want to change the coordinates, you do:
>
>Mark(field, "positions");
>
>Compute ("[a.y * cos(a.x), a.y * sin(a.x)]") ...or whatever the
>conversion is between your coordinate system & cartesian.
>i think this is right assuming that your coordinates are [r,z].
>
>Unmark(field, "positions");
>
>that converts the positions from cylindrical to rectangular.
>
>if your data is scalar, you can leave it. if you have [r,z] vector
>data, then you have to convert the data as well - but you don't
>need the Mark/Unmark, just the same Compute as above.
>
>if you want to do vector operations like you listed, or anything else -
>like add cutting planes, you can do it in the [r,z] coordinate system,
>without any special flags or settings, and then only do the conversion
>to rectangular before you plot. it gives you different answers depending
>on when you do the conversion, and you can choose which one matches
>the physics of your problem best.
>
>think about a cutting plane added in [r,z] space. when you cut at a
>constant z and then convert to [x,y], you get a semi-circle. but if
>you convert to [x,y] first and then cut at a constant y, you get a
>straight line. you might want either or both of these depending on
>your problem.
>
>hope this helps,
>nancy
>
