Does this patent mean I'm not allowed to do adaptive mesh subdivision?
1–10 of 28 posts
Re: Does this patent mean I'm not allowed to do adaptive mesh subdivision?
#2if you are going to worry about violating software patents, you might as well stop coding, because you probably violate 100s or 1000s with anything remotely nontrivial
Re: Does this patent mean I'm not allowed to do adaptive mesh subdivision?
#3if you are going to worry about violating software patents, you might as well stop coding, because you probably violate 100s or 1000s with anything remotely nontrivial
probably just by pressing the power button on your machine. If you expect to have revenue; also expect some of that revenue will be appropriated by lawyers.
Re: Does this patent mean I'm not allowed to do adaptive mesh subdivision?
#4if you are going to worry about violating software patents, you might as well stop coding, because you probably violate 100s or 1000s with anything remotely nontrivial
I'd guess PDE people could claim prior art on this?
Re: Does this patent mean I'm not allowed to do adaptive mesh subdivision?
#5I dunno. Either there's something subtle going on here that I'm not getting, in which you're probably safe unless you're actually using their code; or there is loads of prior art on this idea.
I worked on HyperMesh [1] in the mid-1990s, and while it wasn't my area of the code, we had a heavy focus on automatic mesh generation from geometric surface data. All this patent sounds like is applying the simpler bits of those techniques to meshes which had already been generated. (We had a ton of tools for that as well, but maybe none so automatic as what they are describing here?)
Of course, this being the world of software patents, none of what I've said indicates you cannot get seriously screwed by this patent...
[1] http://www.altairhyperworks.com/Product,7,HyperMesh.aspx
Re: Does this patent mean I'm not allowed to do adaptive mesh subdivision?
#6Also, patent titles and descriptions can be very general, even if the actual patentable claim is very specific.
For example, I am on a European Patent titled "Method and Apparatus for producing a printing pattern" No, we did not patent "producing a printing pattern", it was one very specific method of doing so. (I use "we" very broadly, I didn't actually file the patent, they took a vaguely worded e-mail from me and copy-pasted it into the application, but that's another topic...)
Re: Does this patent mean I'm not allowed to do adaptive mesh subdivision?
#7Here's the first claim, slightly reworded to be easier to read:
> 1. A computer-based system for smoothing a mesh model of a three-dimensional object or surface,
the mesh model having a set of connected faces, each face being a geometric shape with a predetermined number of vertices and a corresponding number of edges, each edge connecting two of the predetermined number of vertices, each vertex being a data point describing the object or surface ... comprising:
a subdivision [function] to check a first edge against a predetermined subdivision criteria;
an insertion [function] to insert a new subdivision point along the first edge, the new subdivision point being inserted depending on whether or not the first edge meets the predetermined subdivision criteria; and
an extrusion [function] to extrude the new subdivision point so that it lies on a curve, wherein the curve passes through the vertices connected by the first edge orthogonal to normal vectors at the vertices.
(Some "subdivision criteria" are covered in dependent claims.)
I'm not a graphics / computational geometry guy, so I don't know how common / standard this is. The last step about extrusion through a curve looks somewhat narrow to my non-ordinary skilled eyes. Are these steps what you had in mind?
The 4th claim, however, is a doozy. As far as I can tell, it essentially covers recursively applying subdivision to only a selected subset of faces in a mesh:
> 4. A method for refining a mesh model of a three-dimensional object or shape, the mesh model including a plurality of first geometric shapes, comprising the steps of:
selecting at least one of the first geometric shapes;
recursively subdividing each of the selected first geometric shapes into a plurality of second geometric shapes; and
displaying each of the plurality of second geometric shapes.
Again, not familiar with the field, but this one sounds incredibly broad.
This was filed in 1999, so it has a few more years left.
Re: Does this patent mean I'm not allowed to do adaptive mesh subdivision?
#8if you are going to worry about violating software patents, you might as well stop coding, because you probably violate 100s or 1000s with anything remotely nontrivial
Thanks for your input. I'm a contractor and i've been asked to implement a system for which dynamic mesh subdivision seems like the best solution.. additionally i'd like to subdivide only small regions of the target mesh. I'm wondering if I should raise the question of this patent with my employer... the patent does seem overly broad... just wondering what to do from an ethical/professional standpoint..
There are lots of apps out there that do various forms of this.. dx11 has hardware implementations of it.. but if I implement a software version that implements a simple midpoint subdivision/smoothing, where I smooth the midpoint vertex by some amount based on neighboring edges.. seems to get close to some of the language in that patent.. I'll probably mention it to my employer just in case.
Re: Does this patent mean I'm not allowed to do adaptive mesh subdivision?
#9I dunno. Either there's something subtle going on here that I'm not getting, in which you're probably safe unless you're actually using their code; or there is loads of prior art on this idea. I worked on HyperMesh [1] in the mid-1990s, and while it wasn't my area of the code, we had a heavy focus on automatic mesh generation from geometric surface data. All this patent sounds like is applying the simpler bits of tho…
Thanks for posting.. you mentioned working on similar stuff in the mid 90s.. could that qualify as prior art since the patent was issued around 99?
Re: Does this patent mean I'm not allowed to do adaptive mesh subdivision?
#10I dunno. Either there's something subtle going on here that I'm not getting, in which you're probably safe unless you're actually using their code; or there is loads of prior art on this idea. I worked on HyperMesh [1] in the mid-1990s, and while it wasn't my area of the code, we had a heavy focus on automatic mesh generation from geometric surface data. All this patent sounds like is applying the simpler bits of tho…
Thanks for posting.. you mentioned working on similar stuff in the mid 90s.. could that qualify as prior art since the patent was issued around 99?