CompuServe Thread

#different UCS

4 messages in this thread
#21654From: Tony Tanzillo [LISP TM]Nov 23, 1991 6:04 PM
Dietmar – That's a neat idea, but can't you do this right now? The UCS table contains the information needed to find the Z-axis vector of the UCS, and the (trans) function will accept that as well. I've never tried it, but it should be possible, requiring a little matrix math. -TonyT.
#21880From: Dietmar RudolphNov 25, 1991 12:11 PM
Tony, well, you're quiet right, but … The UCS table contains the X and Y directions but not the Z direction. OK, you can calculate that, but I'm too lazy to do that. Always thought that calculations are something for computers not for people <g>. BTW I doubt that it will be easier/faster to scan the UCS table, calculate the Z axis and put that into (trans) than to (trans) into WCS first, change UCSes and (trans) back. So I make this a wishlist item. Dietmar
#21958From: Tony Tanzillo [LISP TM]Nov 25, 1991 8:54 PM
Dietmar: > But I doubt that it would be easier/faster to scan the UCS table, > Calculate the Z-axis, and put that into (trans), then to trans into > the WCS first, change UCSes and (trans) back. I don't understand, Dietmar, if you do this using the UCS command, then what do you think AutoCAD does in order to determine the new UCS? It has to go thru the same steps (look up the saved UCS, compute the normal or Z- axis vector, and then compute the internal transformation matrix that it uses to compute coordinates in the UCS). If you want to make a wish for (trans) to accept the name of a saved UCS, then it too, must lookup the UCS, compute the matrix, and do the transform, but I don't know how much faster AutoCAD can do it, as opposed to an ADS program (of course, LISP is just plain slow at this, which is one argument for the feature). I'm not opposed to it, I just think its easy enough to do right now, and I thought I'd bring that to light. -TonyT.
#22030From: Dietmar RudolphNov 26, 1991 2:35 AM
You're right, an ADS function ought to perform this task as fast as ACAD does, or even faster <g>. I was talking to LISP, so performance and ease-of-use are still a matter of fact. On the other hand I'm not so enthusiastic about ADS because of some reasons: – There are still versions of ACAD that don't support ADS and they will for some years. – You can port AutoLISP code across platforms in a snap. This doesn't hold for ADS code (at least you'll need another compiler and linker) – There are gigabytes of tested and debugged LISP code. Why rewrite that? – Performance is not a key issue in complex design tasks. E.g. an application of mine (which I call typical) runs 2 minutes with user interaction disabled. In the standard case with user interaction it runs about 20 minutes since the user has to make some design decisions he has to think about. So there's no need to reduce the program's running time to 1 minute and the total time to 19 minutes. – Last point: I'll wait for ADS support of a high level language like LISP instead of low level languages like C or FORTRAN. Dietmar