Version: 5-jul-94 Dear fellow particle pusher: The questionnaire you wil find below was designed to get an idea how different particle pushers describe their persistent data. The idea is that we can use this questionnaire to describe a reasonable common dataset (not dataformat!) that can then be used to construct an interchange and archive format to enable different researchers (not just the experts) can more easily obtain datasets/simulations from their collegues, and use their own software to analyse the results. NASA has expressed interest in storing such data, although they may be keen to see this happen in FITS format. Observers have done this kind of exchange and archiving for years using the FITS format, and there is no reason we should not (although not neccesarely FITS). ==> Anybody who would like to be involved in this process is encouraged to contact me about that! Your Name : - E-mail : - Software : - Range of 'N': - *** As an example, the answers for NEMO are placed on the right, ------| *** in the last column v Kind of simulations you do (add others if you know about them): ( ) N-body (algorithms don't matter?) + ( ) SPH N-body - ( ) other: - Kind of software you use for analysis: ( ) NEMO + ( ) TIPSY ( ) StarLab ( ) Nora ( ) my own hown brewn (please enter name if it has one): ( ) other (IDL/Matlab/Mathematica/...) The outset: ( ) A simulation result can be described as a series of + snapshots, i.e. each of them at a particular moment in 'time'. (sometimes 'time' isn't quite time, but can be mapped to it) with the following classifications (add more if need be) Each category is detailed with examples below. ( ) 0. single simulation variables, that describe the whole simulation + ( ) 1. single snapshot variables, that describe each snapshot + ( ) 2. particle variables, that describe each particle in a snapshot + ( ) 3. group variables, that describe certain subsets of the particles - in a snapshot (e.g. membership, class etc.) ( ) 4. ... - 0. simulation variables - add your own variables if you have them. ( ) history + ( ) comments/headline + 1. snapshot variables ('header') - add your own variables if you have them. ( ) nbody the number of bodies + ( ) ndim the number of dimensions + ( ) time the time of this snapshot + ( ) coordsys some encoded way how the 'ndim' numbers should be viewed + (e.g. xyz, or zxy, or r-theta-phi etc.) ( ) units ? - ( ) gravc gravitional constant, if relevant - ( ) mass only if all masses the same (see also below) - ( ) softening grav softening (see also below) - ( ) other: 2. particle variables ('data') - add your own variables if you have them Sometimes a particle variable is constant (mass, softening), and may be promoted to a snapshot or even simulation variable. ( ) mass (may also vary as function of time) + ( ) positions NDIM (2 or 3 presumably) - ( ) velocities NDIM (assumed to be d(position)/dt) - i.e. same coordsys - ( ) phase space 2*NDIM + (tag this only if you always carry pos & vel together) ( ) potential + ( ) forces NDIM (assumed to be same coordsys) ( ) jerk NDIM (assumed to be same coordsys) ( ) density - ( ) temparature - ( ) pressure - ( ) smoothinglen sph smoothing length for the previous variables) - ( ) softening grav softening (but see also above) - ( ) luminosity (may also vary as function of time) - ( ) age - ( ) metallicity - ( ) other: ( ) It is assumed that each variable can have it's own type. For example, + mass could be single precision, position and velocity double, but potential and forces in single precision again. etc. 3. particle group variables ('additional table') - add your own ( ) mass array of masses & index pointers for diff mass groups - ( ) other: Please hand back to me, or use: Peter Teuben teuben@astro.umd.edu 301-405-1540 Astronomy Department University of Maryland College Park, MD 20742 USA