[sbml-discuss] Entelos patent on simulations
Upinder S. Bhalla
bhalla at ncbs.res.in
Sun Apr 30 09:28:08 PDT 2006
The simulator GENESIS released in the mid eighties seems to cover most of
the points below. I am sure earlier simulators did the same general
Not that we considered that any of this was remarkable - it was all quite
obvious. But from what I have heard, it is much harder to fight a patent
on grounds of obviousness than on grounds of prior art. And from what I
have heard, both are extremely expensive.
1, 2: GENESIS ran in and with UNIX, and handled system calls nicely.
3: GENESIS simulates neural networks
19, 20, 21: All standard stuff in GENESIS
s/GENESIS/NEURON/g and it would all still hold.
-- Upi Bhalla
On Sat, April 29, 2006 8:00 pm, Bill Denney said:
> Well, I'm not a patent attorney, but this doesn't look like it would stand
> up to me:
> First, all claims stem from claims 1, 2, 3, 19, 20, and 21; so those
> appear to be the only ones that need to be invalidated.
> Claim 1 appears to be utilization of simulation software using the unix
> pipe command-- it would be easy to show that pipe and unix-based
> simulations have been around longer than 2001.
> Claim 2 could be an if block in the shell script that runs the simulation.
> Claim 3 could be a neural network.
> Claim 19 appears to be using a simulation engine, but otherwise the same
> as claim 1. People have written generic simulation engines for years.
> Claim 20 appears to be allowing scripting in the simulation engine. We
> just need to find a simulation engine that allowed scripting before 2001.
> Claim 21 appears to just be a neural network simulator that plots the
> results of each iteration of training.
> Again, I'm not a lawyer, but it appears that this wouldn't be too hard to
> remove. Perhaps we could get help from the EFF
> (http://www.eff.org/patent/) to invalidate this.
> On Fri, 28 Apr 2006, Michael Hucka wrote:
>> Bill Hlavacek pointed this out recently, and at first blush,
>> it sure seems to impact other software out there:
>> Patent Number 6,983,237 awarded January 3, 2006,
>> to Entelos, Inc.
>> "A method of conducting a sequence of linked simulation
>> operations, utilizing a computer-based simulation model,
>> commences with the performance of a first simulation
>> operation to generate an output condition. A further
>> simulation operation, which is defined to sequentially
>> follow the first simulation operation, is then formed
>> utilizing the simulation model. The second simulation
>> operation at least partially and automatically inherits the
>> output condition generated by the first simulation operation
>> as an input condition. In this way, the second simulation
>> operation commences with a configuration captured from a
>> preceding simulation operation as an input
>> condition. Multiple simulation operations may at least
>> partially inherit a configuration from the first simulation
> "No discipline is ever requisite to force attendance upon lectures which
> are really worth the attending."
> -- Adam Smith, "The Wealth of Nations"
More information about the Sbml-discuss