Hey guys,
Ive been working on a procedurally generated spherical hexgrid for a while now and it finally works!
It functions by subdividing an Icosahedron and then taking the dual polyhedron of the result to generate a grid on the icosphere composed of hexagons and pentagons. Each Hexagon or Pentagon tile is its own gameobject, so you can attach your own script to the tiles and have them do what you want.
Ive seen a couple people asking for help with this type of thing and I was wondering if there are enough people interested in this for me to submit my working version on the AssetStore. So far you can dial in the number of subdivisions up to 4 levels, resulting in up to 2560 tiles. There will always be exactly 12 pentagonal tiles for any level of subdivision. Unity tends to freeze up for more than 4 subdivisions, but im working on expanding this.
If anyone else would like to see this up on the asset store, Id be happy to submit it. My projected price is around 5$ for package. Let me know what you guys think!
Cheers
(Sphere with 1 subdivision)
You could easily sell it for a lot more if you include helpers for determining n tiles between tiles, direction of shortest arc between tiles, placement of objects on the tiles, a* (or any) path finding along tiles, and arbitrary camera orbiting. That could all be done by the user anyway though. Instant purchase for me either way.
Hey guys, sorry for being MIA for a while. I didnt realize there was this much interest in this asset!
If yall are still interested, I can put together a package for the asset store. Im thinking a price of around 3$ would be reasonable? Chime in if you would be willing to pay that amount. (This took a TON of work so Id prefer not to make it free)
Thanks again for the support!
Cheers
Great to hear that, Im currently updating it for Unity 5 and Ill submit it to the asset store today. In the mean time, heres a few screenshots from a game I made a while back using the hex sphere
Just resubmitted to the asset store! Should be a couple of days before approval and then its live!
Hope yall enjoy it, thinking of making a slightly more expensive pro version with pathfinding and some more infrastructure built in to allow for quick game creation.
Hey ChicknBoots,
I have purchased it and wanted to make a review. Before doing so I have some complaints/suggestions you could improve to recive a higher rating ;).
You have exported it as a “complete project” which simply overwrites your customers project settings when importing if not unchecked. So it would be better to leave these out from the package entirely when there is no urgent need for it. I find it quite demanding and invasive fur such a simple asset and restricts it from beeing imported into existing projects.
Importing the package into an empty project spawns 5 warnings (assigned but never used) into my console. I’m allergic to packages which do this since I need my console to display important stuff and don’t want to search it in the clutter of “lazy” package creators. So I (and certainly others) would apreciate if my console stays clean when I import your package.
There are no prefabs provided. Sure one can assemble them themselves but why put this inconvenience onto your paying customers? When dealing with a purchased package I expect to provide the functionality in a drag and drop manner and not in “bare bones” where I need to construct everything myself.
When running the example scene I get the error Tag “Tile” is not defined. I guess because I did not import the project settings. This requirement should be mentioned.
In the camera script moving up/down is reverted.
Your code could use some comments. Especially public fields (inspector modifiable) should state what they are used/required for.
Creating a GameObject for every tile is inelegant. And having a mesh collider for each is quite expensive. I thought the whole sphere is one mesh.
If you fix issues 1 to 6 I could offer you a 4 star rating. But in my opinion the asset is only usable for simple projects due to the performance implications of your “many GameObjects with colliders” approach. I thought it could be improved by “terrain generation” and path finding (fe for strategy games) by myself but I’m not sure its suited for that.
Given the price, I think I had much lower expectations than @exiguous , but his thoughtful feedback is great if you want to substantially improve the product. If you do implement his suggestions, I really think the price should be increased. Either way, I’m happy with it.
Regarding exiguous’s comments:
Agreed. If the Project Settings are included to setup the Tags, maybe it should just be mentioned in the read-me that the “Tile” tag has to exist. Or remove the tag reference in the code. I personally won’t be using them.
It’s not really important to me given the price. If it was >$20 then I’d agree. They’re very easy fixes though so maybe it should be fixed.
Would be nice, but see #2.
See #1.
I think the camera is fine. I imagine it like pressing in the direction that you want to “push” the planet away. That said, maybe there should be an option to invert the direction.
Would be nice, but see #2.
I disagree with having just one mesh, but it would be nice to have modes to choose what you want. I modified the script to make each tile have a local space so that I can align surface objects without hassle. That wouldn’t be as easily doable with a single-mesh version. I’m divided on the multiple mesh-colliders though. A sphere-collider would be faster, but it’d be dysfunctional on the lower-resolution versions.
Thanks for the suggestions! I wasnt aware that it would overwrite your project settings, Im sorry about that and it will be one of the first things I fix.
I can easily update the package to have a prefab, although its as simple as attaching the main script to an empty and assigning 2 materials. Youre right though, it would be nice to have it drag and drop ready.
Im going to get around using the Tile tag so you wont have to import those settings or define it yourself.
In terms of creating a gameobject for every tile, this was the main point of the package: This allows you to change the materials and attributes of each tile independently and makes it much more flexible. If the entire sphere was a single mesh, changing a tiles color would require altering the material of the entire sphere each time. I think its a bit more user friendly to be able to deal with any tile you want individually instead of going through the sphere’s script. I was considering something similar to your suggestion, however, by having the planet be a single mesh and each tile being a submesh.
I was going to put in an option to choose between a mesh collider or something simpler in a future release, but I didnt think the mesh colliders would be too bad considering they have only 5 or 6 vertices.
In future releases, I do plan on implementing pathfinding as well as some templates for unit movement.
Thanks again for the reply! Ill be working on these issues and releasing an update as soon as possible.
Cheers
Im honored you think so! As this was an initial release with fairly limited capabilities, I set the price really low. I am planning to raise the price to 10$ after the major updates are finished.
