I think there are a few here who have used Papervision, right?

A client wants to make an interface where by a user can create in 3D a pipe assembly, by choosing the various components, and perhaps drag-and-dropping them into place. Pipe bits, u-bends, taps, valves, etc, that sort of thing is what they can choose.

So they build up this assembly, then the details of all that get sent back to the server to generate a quote.

So I'm knew to 3D work, except for very rudimentary stuff.

Could papervision produce such a thing do you think? I'm not looking for a solution here, just ideas and opinions. I was planning on looking into papervision for myself anyway, just hadn't got around to it yet, I've only just made the move to AS3.

Cheers for any thoughts.

you could do something like this in PV.... but.... it will be a bitch to make it usable and useful.

I suggest doing some preliminary discovery and paper prototyping to actually work out how the user is going to build 3D pipe assemblies. What is the context or their work space? are they able to define other (non-pipe) objects such as walls, ground etc? how does one navigate this space? how will you address clipping etc? what are you tolerances for accuracy?

If you can answer all those questions and more then 3D is possibly useful. If you want to do it online then PV is a reasonable option.

FWIW - I once built a 2D (top view) irrigation planner that allowed a user to create their property and then define wet areas. It would then auto create an irrigation system and allow the user to modify it using existing parts. It would then create a recipel lst of items needed that could be taken to a hardware store. It took about 5 months of serious development and that was for a 2D application.

well we don't have many details yet, just trying to get a feel for what could be done.

We've sent back what I think is a high quote to cover most needs, 6 weeks my time, plus other time for design and 3D image creation. I just have to make it work, so hopefully that's enough, but I won't really know (and the quote's not fixed) until we really spec it out.

What we have in mind for now is that there are about 15 pipe components to choose from, we don't know how many configurations that will allow yet, the user chooses those, then they set what connects to what, then I'll draw all that in 3D so they can spin the combination around in 3D - there'll be no context, so no walls or floor or whatever. I also suggested that we don't attempt to do the drag and drop in 3D, but that we handle that separately in a 2D interface. Also said that I'm assuming that the connections between parts are fairly simple.

Anything else they can pay extra for.

In the meantime, I'll try to find time to poke around PV and see what's what, I want to learn how to use it even if this job doesn't happen, just cos it looks like fun.

Originally posted by: DontBogartMe I also suggested that we don't attempt to do the drag and drop in 3D, but that we handle that separately in a 2D interface. Good call. I would imagine that it would be much more intuitive to use a combination of orthographic projection and a 3D render window. You could do some really lovely filter work with PV at fixed angles and perspectives rendered down to bitmaps. Fun.

Originally posted by: DontBogartMe In the meantime, I'll try to find time to poke around PV and see what's what, I want to learn how to use it even if this job doesn't happen, just cos it looks like fun. I've not spent a lot of time with PV but it's well worth learning... especially if you can get paid to do it.

There has been some experimentation of drag and drop on the list might be worth a nabble search. If you stick to horizontal plane movement and maybe you could use the scroller /up/down arrows for height and left/right arrow for rotating the whole scene about the centre and +- for zoom, but I guess there are a lot of approaches and you would need to try them, but might be possible to get something working with drag drop in 3d, try tests with planes or cubes.

Recently I worked on a blue screen video site with a friend where essentially it is 2d (but 3d all people on a single plane), mouse did rotation and you tween to different people by clicking on them or the static 2d people pos map, much simpler nav but worked well. Getting the collada and optimising the material meshes might be the hardest aspect, although yet to do much with collada.