 | | Re: [LDAO] Editor Feedback requested
|
|
(...) OK. This is shorting the process, and completely unrealistic given my current goals and contraints, but I imagine a very good approach for a modeling interface would be like the following. Make the UI look like a set of instructions. Each step (...) (26 years ago, 17-Nov-98, to lugnet.cad)
|
| |
| | Re: [LDAO] Editor Feedback requested
|
|
(...) Problem I would have with this is that I don't always (actually hardly ever) insert steps in the model as I build it. I usually enter in the whole thing and then figure out where the steps should go. this interface paradigm would be (...) (26 years ago, 17-Nov-98, to lugnet.cad)
|
| |
| | Re: [LDAO] Editor Feedback requested
|
|
(...) Does that mean the editor should auto-insert steps? Steve (26 years ago, 18-Nov-98, to lugnet.cad)
|
| |
| | Re: [LDAO] Editor Feedback requested
|
|
(...) Karim Nassar commented: (...) And then Steve asked: (...) I don't think so. If you want to use the instructions paradigm, then the proper solution is most likely to let the user have an option for handling every line as a separate step. It (...) (26 years ago, 18-Nov-98, to lugnet.cad)
|
| |
| | Re: [LDAO] Editor Feedback requested
|
|
(...) Wow. That could mean hundreds of step-images, for large models at least. (...) True. I can think of two approaches of the top of my head (use math to do 3D hit-testing, or build a part-bitmap during rendering, and lookup the clicked pixel in (...) (26 years ago, 19-Nov-98, to lugnet.cad)
|
| |
| | Re: [LDAO] Editor Feedback requested
|
|
(...) That's why I'd hesitate to use the "instructions" paradigm... I'm quite partial to the List/line# interface of LEdit. IMHO it is intuitive and easy to visualize. It seems to me that creating instructions sheets are the end result... so in (...) (26 years ago, 19-Nov-98, to lugnet.cad)
|
| |
| | Re: [LDAO] Editor Feedback requested
|
|
(...) Yes, but I would probably not keep more images at a time than what's needed to fill the screen/window. (...) I was thinking of a part-bitmap. I think 3D hit-testing is slower, even counting the time to make the part-bitmap. | Now, can someone (...) (26 years ago, 20-Nov-98, to lugnet.cad)
|
| |
| | Re: [LDAO] Editor Feedback requested
|
|
(...) Right. Or keep just a few more off-screen images, for interactive performance. If there was a set of maps (color, z-buffer and part) for each element, these could be quickly combined into the display images. (...) Probably. There's not a large (...) (26 years ago, 20-Nov-98, to lugnet.cad)
|
| |
| | (canceled)
|
|
|
| |
| | Re: [LDAO] Editor Feedback requested
|
|
(...) You seem to be reading my mind. :-) (...) Yes, but we would have to add information about how the parts stick together to the files. Play well, Jacob ---...--- -- E-mail: Jacob.Sparre.Andersen@risoe.dk -- -- Web...: <URL:(URL) -- ---...--- (...) (26 years ago, 20-Nov-98, to lugnet.cad)
|
| |
| | Re: [LDAO] Editor Feedback requested
|
|
One thing that would be easy to do in ldlite is store a mapping from each pixel to the line number of the top-level .dat file. With a (as-yet-non-existant) tight integration between LDAO and LDlite, a mouse click in the ldlite window could cause the (...) (26 years ago, 20-Nov-98, to lugnet.cad)
|
| |
| | Re: [LDAO] Editor Feedback requested
|
|
(...) Having such piece-based local maps would also help speed up re-rendering the current step's image when a part is modified. :) Especially if we have that optimized 3D hit-testing logic. Because we'd only want to rebuild the invalidated section (...) (26 years ago, 20-Nov-98, to lugnet.cad)
|
| |
| | Re: [LDAO] Editor Feedback requested
|
|
(...) Yes, that's what Jacob and I are saying. (...) You said it! (...) Could the LDLite/LDAO combo be intelligent enough to know when one or pieces has been hidden, and auto-insert a step? If the user has enabled such behavior, of course. (...) (...) (26 years ago, 20-Nov-98, to lugnet.cad)
|
| |
| | Re: [LDAO] Editor Feedback requested
|
|
(...) I'd say there are infinitely many ways to attach two parts - rotations and sliding a plate between a row of studs are two infinites. I don't know how you'd handle this- maybe a snap connection with an override capability? (...) (26 years ago, 20-Nov-98, to lugnet.cad)
|
| |
| | Re: [LDAO] Editor Feedback requested
|
|
(...) Make the "hotspot" for the connection a line, and allow the pieces to slide along the line. For rods (like flagpoles), the "hotspot" would also allow rotation. Some sliding connections would allow pieces to slide off the end (like an axle (...) (26 years ago, 23-Nov-98, to lugnet.cad)
|
| |
| | Re: [LDAO] Editor Feedback requested
|
|
(...) In the above, I meant Mr. *McDonald*. Apologies to both Toms. Steve (26 years ago, 23-Nov-98, to lugnet.cad)
|
