![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Oh, look. I have a patent. Yay.
Oct. 23rd, 2012 10:27 amAs I've mentioned before, I've written a fair number of patents over the years. I hold the US Patent system in extreme contempt (at least the software side of things -- IMO, software patents do enormous damage to the industry with no real benefit), but it's part of how the game is played. Venture-backed startups tend to apply for patents whether they want to or not, because they wind up with some monetary value.
The nice thing about startups, though, is that most don't survive long enough for the damned patent to actually get *granted*, and they tend not to be pursued much afterwards. But I was always sure that Memento would have legs, and the consequence of that is that the first of the patents I helped write there has actually been granted.
(Of course, I found out about this because today I received flyers from not one but *two* competing companies, each trying to get me to spend hundreds of dollars on a mounted plaque commemorating my patent. Not exactly likely.)
Still, for the curious, it's patent number 8,290,845: "System and Method for Presenting Quasi-Periodic Activity". The hell of it is, the patent is entirely valid by the numbers -- it's an innovative and rather clever way of illustrating repeating behaviour in a bank account. One of the problems in check fraud detection is "false positives" -- most of the old systems produce literally *hundreds* of false alarms for every real fraud they detect. We spent a huge amount of work reducing that through many techniques.
Quasi-Periodic Activity (the math team's term) is part of that: if a check is part of a recurring pattern of activity, it's much less likely to be fraudulent. (The "Quasi" part is because much of the repeating activity isn't necessarily exact -- we want to catch "about $82 about every four weeks for the phone bill" and things like that, which made the programming for the math team rather more entertaining.) The important part is really *detecting* that, but we also needed to *display* that to the analyst, so that they would know why we're alerting this check but not that one; the display is really what the patent is on, even though it's the less important part in the grand scheme of things.
The patent is, as usual, totally opaque -- part of why I hate patents is that they take lovely algorithms and turn them into incomprehensible gobbledygook to satify the Patent Office. But the pictures are pretty (if a lot less pretty in black-and-white than the original color), and show you what I spent a couple of years building at Memento. These pictures are just the design, but the actual implementations looked extremely similar -- part of why I built everything in Flash was so that I could take the ideal designs and follow them really faithfully. (Which used to be very hard in HTML.)
The truth is, I consider myself relatively minor here. Greg (our Chief Mathematician) came up with the core ideas, and Katy (our UI/UX designer) came up with the actual look-and-feel; I just *built* the thing so it would actually work the way they wanted it. But that's still important from a patent point of view, so there I am on it. (Not that I actually *own* it in any meaningful sense -- like most such work, the actual ownership got assigned to the company long ago.)
All of which does leave me wondering about the state of the *other* four patents I worked on there...
The nice thing about startups, though, is that most don't survive long enough for the damned patent to actually get *granted*, and they tend not to be pursued much afterwards. But I was always sure that Memento would have legs, and the consequence of that is that the first of the patents I helped write there has actually been granted.
(Of course, I found out about this because today I received flyers from not one but *two* competing companies, each trying to get me to spend hundreds of dollars on a mounted plaque commemorating my patent. Not exactly likely.)
Still, for the curious, it's patent number 8,290,845: "System and Method for Presenting Quasi-Periodic Activity". The hell of it is, the patent is entirely valid by the numbers -- it's an innovative and rather clever way of illustrating repeating behaviour in a bank account. One of the problems in check fraud detection is "false positives" -- most of the old systems produce literally *hundreds* of false alarms for every real fraud they detect. We spent a huge amount of work reducing that through many techniques.
Quasi-Periodic Activity (the math team's term) is part of that: if a check is part of a recurring pattern of activity, it's much less likely to be fraudulent. (The "Quasi" part is because much of the repeating activity isn't necessarily exact -- we want to catch "about $82 about every four weeks for the phone bill" and things like that, which made the programming for the math team rather more entertaining.) The important part is really *detecting* that, but we also needed to *display* that to the analyst, so that they would know why we're alerting this check but not that one; the display is really what the patent is on, even though it's the less important part in the grand scheme of things.
The patent is, as usual, totally opaque -- part of why I hate patents is that they take lovely algorithms and turn them into incomprehensible gobbledygook to satify the Patent Office. But the pictures are pretty (if a lot less pretty in black-and-white than the original color), and show you what I spent a couple of years building at Memento. These pictures are just the design, but the actual implementations looked extremely similar -- part of why I built everything in Flash was so that I could take the ideal designs and follow them really faithfully. (Which used to be very hard in HTML.)
The truth is, I consider myself relatively minor here. Greg (our Chief Mathematician) came up with the core ideas, and Katy (our UI/UX designer) came up with the actual look-and-feel; I just *built* the thing so it would actually work the way they wanted it. But that's still important from a patent point of view, so there I am on it. (Not that I actually *own* it in any meaningful sense -- like most such work, the actual ownership got assigned to the company long ago.)
All of which does leave me wondering about the state of the *other* four patents I worked on there...
