## [Opensg-users] pfDCS and some thoughts about it ...

 [Opensg-users] pfDCS and some thoughts about it ... From: Mathias Gumz - 2002-04-17 21:12:35 ```hello opensg-list, i a mail some days (or weeks) ago i told you about my problem and the=20 "missing" feature of some "dynamic nodes" or something like this. now my suggestions about that: its right, you could play around with quaternions and add these as the=20 transformation core to a node and then you can move this node with the=20 special transformation. the real problem is, that i have 3 separate transformations (or perhaps m= ore)=20 i have to keep separate from each other. and i dont want to use a lot of = node=20 just to apply these different transformations, coz i find the concept of=20 opensg with the one node per position in space very usefull and well=20 designed. so i think, a good method to add something like pfDCS could be a containe= r of=20 transformations (lets call it .. list ? ), where you can append other,=20 sperate transformations to the node. the transformations would stay separ= ated=20 and i think the loss of performance for traversing the list will be very = low. so, what do you think about it ? regards, mathias ----------------------------------------------=20 =20 mathias gumz =20 mail: gumz@... =20 http: http://www.wirrehomepage.de=20 icq : 57869799 =20 ---------------------------------------------- ```

 [Opensg-users] pfDCS and some thoughts about it ... From: Mathias Gumz - 2002-04-17 21:12:35 ```hello opensg-list, i a mail some days (or weeks) ago i told you about my problem and the=20 "missing" feature of some "dynamic nodes" or something like this. now my suggestions about that: its right, you could play around with quaternions and add these as the=20 transformation core to a node and then you can move this node with the=20 special transformation. the real problem is, that i have 3 separate transformations (or perhaps m= ore)=20 i have to keep separate from each other. and i dont want to use a lot of = node=20 just to apply these different transformations, coz i find the concept of=20 opensg with the one node per position in space very usefull and well=20 designed. so i think, a good method to add something like pfDCS could be a containe= r of=20 transformations (lets call it .. list ? ), where you can append other,=20 sperate transformations to the node. the transformations would stay separ= ated=20 and i think the loss of performance for traversing the list will be very = low. so, what do you think about it ? regards, mathias ----------------------------------------------=20 =20 mathias gumz =20 mail: gumz@... =20 http: http://www.wirrehomepage.de=20 icq : 57869799 =20 ---------------------------------------------- ```
 Re: [Opensg-users] pfDCS and some thoughts about it ... From: Gerrit Voss - 2002-04-18 06:04:40 ```Hi, Mathias Gumz wrote: > hello opensg-list, > > > i a mail some days (or weeks) ago i told you about my problem and the > "missing" feature of some "dynamic nodes" or something like this. > > now my suggestions about that: > > its right, you could play around with quaternions and add these as the > transformation core to a node and then you can move this node with the > special transformation. > > the real problem is, that i have 3 separate transformations (or perhaps more) > i have to keep separate from each other. and i dont want to use a lot of node > just to apply these different transformations, coz i find the concept of > opensg with the one node per position in space very usefull and well > designed. > > so i think, a good method to add something like pfDCS could be a container of > transformations (lets call it .. list ? ), where you can append other, > sperate transformations to the node. the transformations would stay separated > and i think the loss of performance for traversing the list will be very low. > > so, what do you think about it ? > just to be sure, you are talking about multiple full transformations, aka 4x4 matrices, and not about storing for example the translation seperate from the rotation, so that setting the translation does not override the rotation ?. If this is the case I do not see any problems with the addition of an additional Tranfrom core which is able to store multiple matrices which are accumulated at draw time. Gerrit ```