@Gilbert, we could also let a CoordSystem have a motion, with something like system_a.set_vel(translational = ..., angular = ....) <- this would be with respect to some system defined in that frame only. Then coordinates of this system, when expressed in some other system, would be functions of coordinate variables AND the 'time' variable of the global reference frame.
On Wed, Jun 5, 2013 at 1:57 PM, Gilbert Gede <[email protected]> wrote: > Sachin, > I like where you are going with this. If I'm interpreting it correctly, > each CoordinateSystem has to be attached to a ReferenceFrame, and is fixed > (although possibly rotated and/or translated upon coordinate system > definition) with respect to that ReferenceFrame? Prasoon, Stefan, others - > what are your thoughts on this? > > I think there might be some better/easier ways to access basis vectors and > some other issues, but that discussion can come after a consensus is > reached on the CoordinateSystem/ReferenceFrame distinctions. > > -Gilbert > > > On Wed, Jun 5, 2013 at 12:21 AM, Sachin Joglekar > <[email protected]>wrote: > >> @Stefan : I would recommend having a separate class for ScalarFields. >> Even I wasn't sure of the need for this till yesterday, when we came across >> the problem of how the user would define a scalar field in any coordinate >> system he wants(which is not the global frame). In such cases, I propose >> something like the following- >> rho = ScalarField(6*x**2*y, c_rect1) >> rho.express(c_sph) >> >> If we find a better way to this this I am all for it, but for now the >> above seems elegant and convenient. >> >> In any case, I tried my hand at expressing a few of the steps put forward >> by Stefan in a mock SymPy session. I would request all of you to have a >> look at it and express your views (please elaborate on the reasons for any >> editing if done in the current code). It is a WIP obviously. >> The wiki link - https://github.com/sympy/sympy/wiki/Vectors-EM-framework >> >> >> >> >> On Wed, Jun 5, 2013 at 4:16 AM, Stefan Krastanov < >> [email protected]> wrote: >> >>> Here is a quick summary from today: >>> >>> - probably scalar fields will be represented simply by SymPy expressions >>> where some of the symbols will have special meaning (the coordinates) >>> - probably vectors will be represented like in mechanics (one object, >>> not necessarily a sympy expression) >>> >>> - using reference systems translated and rotated with respect to each >>> other is rather unclear at the moment: before continuing with the irc >>> meetings I would suggest that the students provide a _nice_ wiki presenting >>> the answers to the questions in the previous thread and also the following >>> questions: >>> >>> Bellow is the question in "mathematical" terms. Transform it in whatever >>> way you find appropriate to fit your suggested APIs: >>> >>> - in 3D >>> - I have three points A, B and C. >>> - I use each of them as the zero of three different coordinate systems >>> - The A and B systems are both Cartesian but rotated by theta_AB around >>> axis_AB >>> - The C system is spherical (r, phi, theta). The theta=0 axis is rotated >>> wrt the z axis of A by the Euler angles alpha, beta, gamma >>> - I define a scalar field in A, another scalar field in B and a vector >>> field in C >>> - I want the sum of the scalar fields >>> - I want the gradient of that sum >>> - I want the convective derivative of the vector field from C wrt the >>> gradient from the question above >>> - I want to express the entities from the above 4 question in each of >>> the three coordinate systems. >>> - For all this please explicitly choose some fields for the examples and >>> calculate the expected results by hand (and add them to the example session >>> as mock results). >>> >>> I think that this will really stress test the suggested API. The only >>> thing missing is the time dependence needed in mechanics. I strongly >>> suggest that we first finish the considerations above before continuing. >>> >>> @Prasoon and Sachin, when will you be able to provide a detailed wiki >>> page with an example session for what is asked here? There is really no >>> need to hurry (officially GSoC has not started yet) so please take your >>> time (a week?). >>> >>> Stefan >>> >>> >>> On 4 June 2013 01:12, Aaron Meurer <[email protected]> wrote: >>> >>>> The discussion was at http://piratepad.net/KBviCWUlA3. >>>> >>>> I'm curious what you think of this kind of discussion, as opposed to >>>> IRC. Google docs is also an option (it has a chat). I think the >>>> downside is that unlike IRC, which is logged at >>>> http://colabti.org/irclogger/irclogger_logs/sympy, it's a little >>>> harder to search through these discussions afterwords. >>>> >>>> Aaron Meurer >>>> >>>> >>>> On Mon, Jun 3, 2013 at 4:16 PM, Stefan Krastanov >>>> <[email protected]> wrote: >>>> > Today we had the first discussion with Prasoon and Sachin about their >>>> > projects. >>>> > >>>> > We did not progress much but at least we outlined the two general >>>> approaches >>>> > that we can use for these modules (specifically for creating vector >>>> fields). >>>> > I will give them somewhat arbitrary names here: >>>> > >>>> > - the `mechanics` way - having a Vector class that keeps all the >>>> information >>>> > about the field and it is not part of expression trees in the way >>>> Basic and >>>> > Expr are. For instance Vector(something along >>>> cartesian.x)+Vector(something >>>> > along spherical.r) will result in Vector(complex internal structure). >>>> > >>>> > - the `diffgeom` way - having base/unit vectors and building all the >>>> rest in >>>> > terms of their linear combinations (all this expressed as sympy >>>> > expressions). >>>> > >>>> > >>>> > >>>> > Prasoon and Sachin did not have the time to look at the example >>>> problem that >>>> > was given in the previous email yet (no harm done there, there is >>>> still some >>>> > time before the official starting date). Probably this will be the >>>> subject >>>> > of our next discussion. >>>> > >>>> > The next discussion was scheduled for tomorrow. After that I suggest >>>> that we >>>> > keep most of the discussions to the mailing list and the gihub wiki >>>> and meet >>>> > on irc / realtime wikis / google docs / etc once a week. >>>> > >>>> > Stefan >>>> > >>>> > -- >>>> > You received this message because you are subscribed to the Google >>>> Groups >>>> > "sympy" group. >>>> > To unsubscribe from this group and stop receiving emails from it, >>>> send an >>>> > email to [email protected]. >>>> > To post to this group, send email to [email protected]. >>>> > Visit this group at http://groups.google.com/group/sympy?hl=en-US. >>>> > For more options, visit https://groups.google.com/groups/opt_out. >>>> > >>>> > >>>> >>>> -- >>>> You received this message because you are subscribed to the Google >>>> Groups "sympy" group. >>>> To unsubscribe from this group and stop receiving emails from it, send >>>> an email to [email protected]. >>>> To post to this group, send email to [email protected]. >>>> Visit this group at http://groups.google.com/group/sympy?hl=en-US. >>>> For more options, visit https://groups.google.com/groups/opt_out. >>>> >>>> >>>> >>> >> > -- You received this message because you are subscribed to the Google Groups "sympy" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/sympy?hl=en-US. For more options, visit https://groups.google.com/groups/opt_out.
