Building and arched door way

Is there a means to build an arched door (entrance) way using the SL prim collection such that it can fit into a 5 x 10m brick wall. I can make the arch and the adjacent walls, but it's the adjacent wall above the arch with the missing quarter circles that are tough. Any help would be appreciated? Thank you.

The way that the arch is typically done is to take a 5m by 5m wall section and profile cut it to 12.5% to 67.5% (it's ok if the editor chops that to 13% and 68%). Then make it hollow and set the hollow shape to a circle. The z-axis of the wall prim (a box prim) needs to be pointing outward for this to work

You should end up with a 5m by 2.5m rectangle with a half-circle bite taken out of the 5m side. Then you build the side pieces to match.

Brilliant!!!!!!!!!!!!

You can also build an arch doorway using only 2 prims.
First build the top part of the doorway as described above.
After that copy that down.
Rotate the lower piece upside down.
Change the hollow back to a square.
Lower the bottom prim below floor level until the bottom cross member disappears beneath the floor.
Stretch the bottom sides up to meet the top side walls.
Keep adjusting the sides till you get it the height you want.
The bottom part shouldn't matter if it's in the ground a little.
And as a plus you've got a much neater door opening and you saved 1 prim.

As for the door itself, if you are putting one in the arch, there are two ways to go. You can make a multi-prim door with a round rop, and texture that. Or you can make a one-prim rectangular door, alpha-texture the front, back, sides and top to show a rounded top, and then insert a hollowed and rounded prim to make the visible top edge, linking that to the door.

I've tried both ways, and the best look came from simply making a rectangular lower door prim and a half-round or quarter-round top door prim, and linking those. With a little effort on textures and repeats, a single texture can cover the faces of the door seamlessly.

Actually, there's yet a third way. I make a one prim rounded double door... (it does rely on alpha though). I'm sure there are other ways as well.

A sculpty door should work well - you can even get the hinge exactly where you want it.

Just keep in mind, a sculpty is 2048 polygons, while a half cube with a half cylinder on top is only 192. That's a difference of almost 11 fold. If you use just a half cube with an alpha texture on it, that's only 84 polys.

I used to talk all the time about how poor texture size management was the biggest source of lag in SL. Now I find myself talking about poly-count management almost as often.

Obviously, one little door will never make or break anyone's FPS, but the principle is very important. Success in building for SL (and for all real time environments) means always balancing appearance, function, and performance concerns with equal weight.

Here's how I'd look it at:

Is it worth spending 11 times as many polygons to save one prim? I would have to say no.

Is it worth spending 11 times as many polygons to avoid using one alpha texture? Again, I'd have to say no.

Is it worth using an alpha texture to save one prim? Mathematically, probably. FPS-wise, maybe. But considering the potential for sorting issues with an alpha'ed door, I'd say it's probably better to use the extra prim.

So, unless I were really, really, REALLY strapped for prims, I'd go with the half cylinder on top of the half cube.

For me, the ability to link a single prim door to the door frame and still have it swing correctly makes it worth it.

Any single-prim door will swing correctly if you just cut it in half. Half-cube doors have been a staple of SL building since day 1. Is there some other way you want a door to swing than on its edge?

Yes, that's my point. However, a cube + half cylinder isn't a single prim.

No, but a half cube with an alpha texture on it is.

True, it wouldn't have depth over the curved part, but to use 24 times fewer polygons (84 instead of 204, it's worth it in principle, performance wise. The only instance where I'd say it's not would be if you were reasonably certain you'd run into sorting issues with the alpha texture, or if for some reason your texure is so big that having another channel to it would use more graphics horesepower than the polys would.

In other words, if how it swings is the only consideration, spending 11 times the polygons is a poor investment, since there are better performing options available. The weight of that particular function is pretty light; the performance hit is heavier. If, however, you add the weight of another issue to the scale, such as alpha sorting, then the sculpty might become worth it after all.

My only point is everyone really needs to start considering all aspects of balance a lot more seriously, than most people do now. We're living in a time where LL is going around nerfing everyone's LOD because too many people are abusing sculpties, causing lag. You might well be balancing perfectly well already, Stephen, for all I know, but many people reading this likely are not yet. I think it's worth spending a little time on this topic whenever the subject of "well, you could do that with a sculpty" comes up is all.

Using an alpha texture to create the impression of a curved top only works if it's either wafer-thin, or never viewed edge on. Sadly, neither of these are true for a door.

If hierarchical linking, or custom meshes, or lower-polygon sculpties were available, they would all be solutions to this problem. Alphas aren't, in my view.

If increasing lag by a factor of 11 every time you need to make a door seems like a worthwhile trade to you, then I guess there's nothing I can say to convince you it's not. We'll have to agree to disagree. I would just ask that others reading this consider carefully what I've said about balance.

You can spit out all the "if we only on had _____" that you can think of, but in the here and now, we have to work with what we've got. In my opinion, while a door that has thickness on all sides but one might look slightly inconsistent to those who would examine it closely enough to notice, the fact that their frame rate will be high enough to notice it in the first place makes it worthwhile.

And again, it's not that one single door will ever make or break anyone's FPS; I'm speaking in principle here. As with large textures, or resource-heavy scripts, or prim hair, or bling, or anything else, keep your sculpty counts as low as possible if you want to maintain the most lag-free environment you can.

... doesn't anyone cheat anymore?

an arched doorway only needs one prim if it's in an existing wall...the cap which is usually a cut + hollow cube, sized into the walls to make it nice and seamless...

if for instance your doorway is 1m wide, make the arch 1.53m wide (1m/.95hollow)

if texture flicker(from the oeverlap) is an issue move it in or out the tiniest fraction (seems to have more precision than shrinking it's depth and so is less noticable)

multiprim doors can swing correctly with a halfcube cut on the root but only if the door isn't linked to the larger structure (yes you could get close with a position moving door, and link messages, but it gets messy and will not alway update all pieces smoothly together)