Follow Slashdot stories on Twitter

 


Forgot your password?
Close
typodupeerror
×
Patents Games

Unity Patents 'Methods and Apparatuses To Improve the Performance of a Video Game Engine Using An Entity Component System' (twitter.com) 46

Posted by BeauHD from the powerful-patents dept.
slack_justyb writes: Unity has filed a patent with the USPTO for "Methods and apparatuses to improve the performance of a video game engine using an Entity Component System (ECS)."

ECS methods are something that some other open source game engines already use. One example is Bevy for Rust. Some are already commenting on the ramifications of this patent application and indicating that this could be a massive overstep by Unity to attempt to patent something already used by other lesser-known game engines.
This discussion has been archived. No new comments can be posted.

Unity Patents 'Methods and Apparatuses To Improve the Performance of a Video Game Engine Using An Entity Component System' More

Unity Patents 'Methods and Apparatuses To Improve the Performance of a Video Game Engine Using An Entity Component System'

Comments Filter:

  • Issued a patent (Score:3)

    by dpille ( 547949 ) on Tuesday September 07, 2021 @07:58PM (#61773817)
    I dunno about "already in use" when the linked patent issued more than a year ago and was filed in 2018. It may even have foreign priority to a year earlier, but when the summary is so bad to start with, I'm not going to dig into the file wrapper to prove it that much worse.

    • Re:Issued a patent (Score:5, Informative)

      by backslashdot ( 95548 ) on Tuesday September 07, 2021 @09:45PM (#61774073)

      Assuming that is true, having applied for patents, I know that patents are (legally, that is) only supposed to be issued for things that are "nonobvious to one reasonably skilled in the art." Patents applications are not disclosed until 18 months after the application. If others have come up with the same idea independently, it means the invention is an obvious one .. therefore a patent shouldn't have been granted. Nowadays patents are pretty much rubber stamps with maybe one or two token back and forths with your attorney. In fact, the only actual difficult kickback I've gotten from the USPTO is them saying to split up claims into multiple patents because it's too much work for them to read (not even joking, that is what they said almost verbatim.) I don't know if they even google to check whether the patent is original. It seems they take the "let the courts sort it out" attitude.

      Reference on patentability requirements: https://www.uspto.gov/web/offi... [uspto.gov]

      Even scarier is the patent backlog ..some of which dates to the 1980s .. yeah .. suddenly a patent may show up that nobody knew about and then everyone owes backroyalties on something ridiculous like using a trackpad.

      • I'm surprised most by the notion that USPTO employees can read.

      • If others have come up with the same idea independently, it means the invention is an obvious one

        Not necessarily. How if 10 people out of 1000 come up with the same idea, how obvious is it? How about if 6 of those people know each other and communicate about the base technology?

        • Re: (Score:2)

          by Kremmy ( 793693 )
          If knowledge of the base technology is enough for the idea to be obvious then it doesn't qualify for patent protections.

      • Under-funding and under-staffing is one of the bigger forms of corruption in the USA with the usual lies about big government fallaciously broadly applied to where it does not fit. The other side of this is to regulate things in such a way that the departments are overwhelmed by BS which has the same functional result; then you have the problem of political appointments by people who want to destroy the agencies. This is likely where they hire the major assholes who destroy the morale of the staff, many wh

      • If others have come up with the same idea independently, it means the invention is an obvious one .. therefore a patent shouldn't have been granted. Nowadays patents are pretty much rubber stamps with maybe one or two token back and forths with your attorney.

        From my experience, the obviousness test is essentially never applied. The USPTO will look in the existing patent base for prior art, seemingly subscribing to the fantasy that if there is no patented prior art, then no prior art exists.

        That was how it worked when I filed for a patent: the only feedback from the USPTO was to point to patents that were vaguely similar and we easily refuted why they didn't constitute prior art. When I was in a defensive position against a very weak patent, it didn't take muc

    • Also, isn't issuing patents on blatantly obvious things that others have been doing for years the stock in trade of the USPTO? Imagine what would happen to their revenue stream if they only allowed patents on novel, nonobvious things!

    • Entity Component Systems?

      Massive prior art. Codemasters used it a good 15 years ago working on Operation Flashpoint and theres probably earlier precedent than that. There are implementations going back a long way for Unreal. I believe Crytek uses it.

      And anyway, its pretty much just the composition pattern with a data layer over the top anyway.

    • Re: (Score:2)

      by juancn ( 596002 )
      There's prior art at least since 2007. "Operation Flashpoint: Dragon Rising" used an ECS system, even Apple's 2015 GameplayKit used an ECS. Unity did do an impressive demo around 2018, but they were not the first.

  • The abuse of the patent system continues. (Score:4, Insightful)

    by Pierre Pants ( 6554598 ) on Tuesday September 07, 2021 @08:24PM (#61773891)
    and it will continue as long as the corrupt system allows it, because it's massively profitable for them. The amount of ridiculous and shockingly unjustified grants of software patents is insane, as anyone who had to check knows. Hundreds of thousands of patents classified as "computer implemented inventions" are granted every year in the US. Then, everyone is surprised about "patent trolls" and once in a few full moons an anecdotal article about an unjustified software patent shows up when it's related to some company that's popular in geeky circles.
    • The system allows what the people let it allow.
      Want to fix this problem? Quit pumping out citizens with an IQ of 37.

      • Want to fix this problem? Quit pumping out citizens with an IQ of 37.

        We're trying, but the republicans keep shitting on public education as hard and as fast as they can.

        We're trying to fix the problem of nonrepresentative democracy, too, but the republicans keep gerrymandering as hard and as fast as they can, they do about ten times as much of it as democrats.

        In fact this has made people so fucking stupid that they don't realize who's doing it to them

        • What borders on stupid?
          Mexico and Canada.

          I've always lived here; I've seen the war on the education Nixon started and the GOP continues to this day; with only his war on drugs beginning to wind down. Even the educators don't realize who is to blame it's so bad; while the most people seem to realize the stupid drug war; just a bit more realized Afghanistan... after 20 years; so maybe we improved a little? Took 60 to realize the drug created prison state...

        • It's fucking aggravating.
          As a highly paid expert with a pretty fucking well developed portfolio of work, having to deal with every fucking moron who thinks they're an expert these days because they watched some fucking talking head on youtube, or wherever the fuck these dolts get their misinformation, I can't help but to think I'm surrounded by fucking morons.
          And it's getting worse. Worse and worse.
          These fucking dumpster fires with an IQ of potato are so fucking invested in their misinformation that if y

    • "with a computer" isn't the issue here. That can be struck down for not being novel.

      The issue is that the laws were changed a few years ago so it's first-to-file, not first-to-invent.

      So the companies that have the money and knowledge to file patents for everything that they think they can get to stick do ... and there's no 'prior art' that can get it taken away from them.

      But the little guy inventor who doesn't have some team that can write up patents for him? He's the one getting screwed, when someone els

    • The problem is that the "non-obvious" standard is far too low a bar. The more people you have, the more negative effects patents have. Patents work well in systems with small numbers of people and thus prevent blatant copying. As the number of people increases, the chance something would be invented multiple times gets patented increases in order N^2. Whereas the benefits of patents must scale in order N. With harm that scales in O(N^2) (O(N) patents times O(N) people) and benefits that scale O(N) there mu
  • Comment removed based on user account deletion

  • Ignore the clickbait title (Score:4, Insightful)

    by dbrueck ( 1872018 ) on Tuesday September 07, 2021 @10:56PM (#61774217)

    A lot of the titles on articles about this have been pretty click-baity.

    They patented a specific strategy for optimizing a particular approach to implementing ECSs. I'm not saying that makes it a good patent, but comments about how ECSs have been used in such-and-such engine for many years are kinda missing the point.

    • Re: (Score:2)

      by Halo1 ( 136547 )

      A lot of the titles on articles about this have been pretty click-baity.

      They patented a specific strategy for optimizing a particular approach to implementing ECSs.

      Actually, claim 1 patents the storage strategy that is literally described in the original blog post that introduced the concept of ECS [t-machine.org].

      I'm not saying that makes it a good patent, but comments about how ECSs have been used in such-and-such engine for many years are kinda missing the point.

      Not really, since a major advantage of implementing ECS as originally described is that enormously helps with cache locality, so (pretty much?) everyone that uses it implements it like that.

      • A lot of the titles on articles about this have been pretty click-baity.

        They patented a specific strategy for optimizing a particular approach to implementing ECSs.

        Actually, claim 1 patents the storage strategy that is literally described in the original blog post that introduced the concept of ECS [t-machine.org].

        Not really - claim 1 (and the patent) is focused on a specific technique to achieve efficient memory management (basically, grouping an object's components into managed slabs of memory). The blog series you linked to sorta mentions that, but mostly in the section about *other* ideas, calling it more of a "runtime/compile time performance optimization, not a design-feature". And to be clear, I'm not saying I think their claims hold any water at all - what the patent describes seems to be stuff that game devs

        • Re: (Score:2)

          by Halo1 ( 136547 )

          Actually, claim 1 patents the storage strategy that is literally described in the original blog post that introduced the concept of ECS [t-machine.org].

          Not really - claim 1 (and the patent) is focused on a specific technique to achieve efficient memory management (basically, grouping an object's components into managed slabs of memory). The blog series you linked to sorta mentions that,

          The blog post to which I linked literally describes that in the "Tables 6,7,8...N+5" section. This is incidentally the post in the series that describes how you would actually implement an ECS system.

          Not really, since a major advantage of implementing ECS as originally described is that enormously helps with cache locality, so (pretty much?) everyone that uses it implements it like that.

          Hmm, no. What Unity patented is one way to achieve memory efficiency with an ECS (again, I think their claims are flimsy or bogus - the techniques they describe don't seem new), but many game engines don't use any approach like this at all, because for most games on many platforms, it simply doesn't matter - there are far bigger gains to be had optimizing elsewhere. The blog series you linked to actually emphasizes this point - it talks about the Playstation and its relatively tiny amount of memory and the performance cost of a cache miss.

          Have you talked to any gameplay programmers about ECS? I have (I work in a game development company), and while the composition is certainly considered to be a nice-to-have, the things they really salivate about are the performance implications of reducing cache misses and the parallelisation opportunities. Eve

          • Actually, claim 1 patents the storage strategy that is literally described in the original blog post that introduced the concept of ECS [t-machine.org].

            Not really - claim 1 (and the patent) is focused on a specific technique to achieve efficient memory management (basically, grouping an object's components into managed slabs of memory). The blog series you linked to sorta mentions that,

            The blog post to which I linked literally describes that in the "Tables 6,7,8...N+5" section. This is incidentally the post in the series that describes how you would actually implement an ECS system.

            Hehe, I don't know what to tell you, but you're misreading those blog articles, or you're misreading the patent app, or both. Agree to disagree, I guess, since the point is so far afield of why I commented on this story in the first place, but what is being claimed in the patent is related to but not the same thing as those blog posts. Do I think the patent is good? Not particularly. Do I think its claims will hold up? Probably not. Are comments/articles saying that Unity has patented UCS accurate? Not remo

  • Sounds like a blackboarding system used by rule-based expert systems (though not limited to them). In these, you "write down info on a blackboard" you find out about a given entity, and rules examine the data to see if they should activate on it. The data need not even be anything any existing rule is looking for.

    The rule looks for what it needs. Hence it doesn't use predefined complex data structures for a predesigned system for that object.

    It's especially useful to parse out understanding in situations

    • Totally different.

      The basic concept with an entity component system is you have a basic Actor class ie "Player" or "Door" or "Baddie" (aka the "Entity"), and then you compose them out of various other components. So you might have a "render mesh" component, a "make noise" component, a "Follows a behavior tree" component and so on. These aren't inhereted by the Actor, but rather exist as instances on the Actor. Theres usually a data layer that sits over it, that the entity and components are exposed to and c

      • The composition system you're talking about is different from the ECS that is coming up in popularity.

        We've been composing entities in games for years. You have an actor of some kind, and that actor holds a container of components. You instantiate the actor in the world, then you instantiate some components on them, and the combination together can do a bunch of things in the world. But critically, the actor itself is an object and when the game engine ticks, it goes through the list of actors, and tells TH

      • He's right. That's totally equivalent to a blackboard in terms of capabilities. At best there's somewhat different physical data layout, but that's it.
  • Same company who puts in remote lockout code into their product and then extorts their customers for extra licensing money erroneously. This company is not to be trusted to use patents defensively
  • These are no legal ramifications because you can always refer to the prior-art.
  • Yep, no one ever thought of reusing components in software. Wow! what a new and completely unheard of concept.
  • Software patents have time and time again shown that they are issued without thought, mostly common knowledge to specific fields, and cause tons of headaches for engineers that must avoid using a specific compression/decompression technique, graphics algorithm, or some other optimization because the government doesn't realize that engineers in certain fields actually do these "not common knowledge" things all the time
  • If it's already used, how can a patent be provided as there is proof of prior art.

  • Unity would be Ubuntu's shitty graphical shell.
    Unity3D is what "hail corporate" morons use instead of Godot Engine.

Slashdot Top Deals

FORTUNE'S FUN FACTS TO KNOW AND TELL: A black panther is really a leopard that has a solid black coat rather then a spotted one.

Close