On Monday 04 Aug 2008 11:55, Saurabh Nanda wrote:
I believe the problem is with the guidelines given in section 4.11.6.However the clarifications given in 4.11.8 will wean out most useless software patents. Don't you think?
Section 4.11.6: [...]The claim orienting towards a "process/method" should contain a hardware or machine limitation. Technical applicability of the software claimed as a process or method claim, is required to be defined in relation with the particular hardware components. Thus, the "software per se" is differentiated from the software having its technical application in the industry.[...]
Having a hardware or machine limitation (hardware + software bundle) does not in anyway change the fact that a solution could be provided in pure hardware or pure software with a probale infinite number of permutations and combnations of hardware/software elements.
Section 4.11.8: The claims relating to software programme product are nothing but computer programme per se simply expressed on a computer readable storage medium and as such are not allowable. For example, if the new feature comprises a set of instructions (programme) designed to control a known computer to cause it to perform desired operations, without special adoption or modification of its hardware or organization, then no matter whether claimed as "a computer arranged to operate etc" or as "a method of operating a computer", etc., is not patentable and hence excluded from patentability.[...]
Consider a VHDL file that describes a PCI /AGP/USB/other device. Copyright serves the purpose of preventing copying or creating derivative works of this file. Any number of permutations of VHDL elements are likely to produce the same end result silicon (perhaps less effeciently). Now adding a driver to control this piece of VHDL code manifested in some silicon does not in anyway change the fact that one could recreate the same functionality with different VHDL + driver or only silicon without vhdl or only software on a DSP / processor or any number of permutation combinations. Why? because the vhdl is a description which tells the interpreter to produce silicon based on a silicon vendors library description of vhdl circuit elements. The resulting silicon will differ drastically from one silicon vendor to another. Similiarly a piece of software code will produce a different set of binaries depending on libraries created for a particular expression of silicon + compiler.
Should Patenting the physical device created from this VHDL program be allowed. MAYBE this and only this arrangement of circuit elements when not programmed into a silicon device which could be reconfigured to something else with different vhdl. Should patenting code which makes this particular silicon behave in a unique way be allowed?
Patent laws provided for patenting demonstrable Physical devices not mere description of physical devices or combination of description and physical devices or combination of mathematical algos created from descriptions of physical phenomena together with physical devices.
The new amendments are attempting to do precisely this.
The above piece is to make whoever goes there fully aware of the mine field of hardware described in software and combined with yet more software within the device + some piddling code known as drivers to activate the hardware - software combo. Software Defined Radios (Intel, Atheros), GPUs (Nvidia, ATI), Memory (Rambus), Sound devices (loooong list) and the new PITAS laptop ACPI microcontrollers and encryption chips.
Blurring of the line between physical device and their representation as mathematical models, then convinently splitting this representation into hardware + software must be resisted at all costs. ANY patent granted to software is an extremely slippery slope