Bezier lathe - looking for testers

I've developed a new bezier lathe tool that automates the construction of large symmetrical shapes with curved sides, like the onion domes of the Taj Mahal. I'm looking for a few experienced builders who would like to test it out. If you're interested, please IM me.

A little more info:
The tool works similarly to ShapeGen in that it creates a shape by rezzing and combining box prims. It works in two steps. First, the user sets the outline of the shape by adjusting a bezier curve using control handles. Second, the tool creates the shape as if the outline were rotated around the vertical axis. The user can set the number of vertical segments and sides around the cicumference, controlling the smoothness.

I attached a pic of an example.

Mmmmm...I don't have the time (or architectural building experience) to test it, but I'm just dropping in here to say 'thank you' for all the wonderful tools that you develop.

Oooh! Count me in! I'd love to test that, dear!

Hey, is that the tool you dropped on me the other day? I'm sorry to say I haven't had time to play with it yet. That roof in your picture is awesome though, so now I've got to give this thing a whirl. Very cool.

Emily - Thanks, I'm glad you you've found the tools useful.

Ceera - I'll drop you a copy.

Chosen - I gave you the plain old BezierGen before, which just makes curves, but I'll drop the lathe on you today.

I have always wanted to make curves like that building,I would love to test it out.

Alley

Since everyone else is replying here, count me in. I'd send an IM but I'm at work and the boss frowns on game playing. (bugger)

A

Well I was going to IM....I would like to test the item as well, I do some custom work and build a lot so I am always looking for new ways of building things.

I'd love the opportunity to test and provide feedback.

I'd love to test as well.. The roof looks awesome

I have a fairly complex symetrical structure I was requested to build for a resident I lease land to.
I think your tool may fit the bill nicely
Shall I IM you when I am in world tonight or can you drop your tool on Sara Sullivan?

Thanks

I would love to test as well.

Please count me in for testing. Thanks

Count me in if you don't mind. I've been looking to explore some of the more uniquely shaped architecture of some buildings. That would be perfect to help me there, since I can design most of the more simpler shapes.

Just a few questions real quick:

1. Are you wanting feedback on it's stability and how well it works given certain odd or unusual commands?

2. Are you wanting feedback on user-friendliness?

3. How DO you do it? xD

Reinhart - My main motivation for the testing is to find things which don't work right or don't work the way the documentation says they do. If there are things about the way it operates that are problematic for people, I'd be interested in hearing about those too.

I use leprechaun magic in my scripts.

Hi there, I am developing an educational sim called English Village - I would love to test your tool - please send me copy! thanks!

The tool works very good and is very easy to use - Just one suggestion/request - if its possible to make it so when the big structures are built to have an option to have them all Linked

Hi Cadroe,

I am afraid I won't be able to give you feedback at any time soon. Shortly after you sent me your new tool, I lost access to the parcel where i did most of my buiilding practice, and where I had thousands of free prims to play with while designing. My remaining land only has 100 or so prims free, so I can't work on builds that may be prim-heavy. And I will not bother to try to work in a griefer-infested public sandbox.

For a typical 'dome' made with this tool, do you have a rough idea on how prim-heavy it would be?

Out of curiosity, can curves be chosen such that a line perpendicular to the axis of rotation intersect the curve more than once?

Simple examples: a curve shaped like a circle to yield a donut's surface, or perhaps an upside down U, to form a Bundt baking pan shape.

Robustus - I'm going to look into the linking issue. I think I can probably do a compromise, which would be to link the prims in each layer, but not the layers to each other. The problem is that I remove the scripts from the prims after each layer is positioned in order to avoid hurting sim performance (I'd consider shapes with over a hundred prims to be typical with this tool, and that's a lot of listening prims). But maybe I can leave a single script in each linked layer until it's linked to the following layer, in a chain. I haven't looked into it yet, so no promises

Ceera - That's too bad about losing your private sandbox, but I appreciate the interest. You can set the number of vertical segments and sides around the circumference, so how many prims depends on how smooth you want it. But as an example, the cottage roof in the picture in my first post uses 24 vertical segments and it has 8 sides, so that's a 192 prims. I've made glass canopies that were meant not to be smooth that used a few dozen. I also built an Enterprise refit bridge hull that has 12 vertical segments and 24 sides, for 288 prims. Totally crazy, but it looks good

Ged - The definitive answer is kind of maybe. As a test I made a roughly circular curve and it did make a torus. The last set of blocks was inverted, but that might just be a bug, I'll have to look into it. But the tool will let you choose a curve "such that a line perpendicular to the axis of rotation intersect the curve more than once." As far as the Bundt pan, I don't think you could do that in one go because there are two different curves being repeated around the circumference.

Just remember the 1 second per link cost.

You can't really speed it up by having multiple scripts in multiple prims each call the llCreateLink, because each one would require a separate permission request and dialog box to the user, thereby replacing the 1 second per link cost, with a 1 dialog per link hassle.

My skirt generating tool has the option of autolinking, but in my own practice, I find it quicker to manually link.

Your tool sounds very cool -- if I even had time to play with ShapeGen, which I fetched months ago, I'd be wanting to play with the Bezier lathe...

I've played with it a bit, and really like it!

I see the potential problem of having 100+ listeners in the rezzed blocks that have been lathed. It takes a while to select them all and delete

A possible solution might be have the lathe'd object be temporary.. you
have 5 minutes to finalize it. The scripts in the rezzed blocks would self-destruct
after a timeout (making the object permanent), or would respond to a delete
command before that point.

In any case, great work! Thanks for putting this out there.

Another suggestion (although not having seen your code, this might not be relevant):

- on state_entry, have your rezzer object request link permissions

- rather than rezzing only a single prim, have your rezzer object contain a linked object with say 30-40 prims (or more), each scripted to hear linkmessages, but NOT listens (should cut down on sim lag)

- upon rezzing the 30-40-prim linked object, the UUID of the object can be obtained by the rezzer, which can then take only 1 second to link itself (as root) to the 30-40 prim linked object

- send link messages to each prim in turn to perform a llSetPrimitiveParams to set the each individual child prim shape, rotation, and position

In the (near?) future, the 30-40 prim linked object might not even need to have ANY scripts in it, once they implement the new LSL function that allows setting prim params of linked child prims, even in script-less child prims.

This approach might yield you both vastly improved auto-linking speeds, reliance on linkmessages rather than listens for less of a sim lag hit, as well as likely improved rez speed.

Ought to work, I think.

Ged - Thanks for the suggestions. The one second delay does make linking seem like a less attractive option. Your idea of rezzing linked sets and sending link messages among the prims seems feasible. One limitation would be distance. Right now the size of the shape is limited by shout distance (up to 100 meters from the tool), but the maximum distance of a linked set is smaller (I don't know the exact formula). You could get around that by managing multiple linked sets, I suppose, although for a tool which adjusts the shapes "live" (like ShapeGen) it would increase the complexity. Overall it seems like a workable scheme, but for now I'm not prepared to rewrite my block management system, which I use in mutliple tools. Good thinking, though.