On Saturday 27 November 2010 13:08:01 Prasanth Sugathan wrote: > Dear friends, > > The Patent Office has revised the Draft Manual of Patent Office > Practice and Procedure based on comments received, including those > from the FOSS community. The patent office has sought for comments > on the manual to be sent by 4th December, 2010. The draft manual > can be downloaded from > http://patentoffice.nic.in/ipr/patent/Revised_DraftManual_PatentOff >ice_04November2010.pdf * > * > > * * SFLC is working on the comments to be sent to the Patent Office > and we request you to kindly post your suggestions. > > The new draft manual published is a welcome change from the draft > published in 2008, especially in relation to explanation provided > for Section 3(k) of the Patents Act, 1970. The earlier manual > tended to adopt the EPO position by considering the technical > effect of computer programmes. The clauses in the previous version > of the manual have been completely reworked. The new manual clearly > states that “*a computer programme which may work on any general > purpose known computer does not meet the requirement of > patentability*”. > > But there is a lack of clarity in sub clauses (e), (h) and (i) of > clause 08.03.06.10 . > Clause (e) states "If a claim in a patent application is not > directed at a computer programme per se it could be patentable, if > all other patentability conditions are met. This provision thus > necessitates distinguishing computer programmes per se from other > types of inventions that uses or implements computer programmes."
So a claim for a fourier transform that extracts vectors from an image stream* is not directed at a computer programme, but the computer programme implementing such a fourier transform is patentable. * essentially execution of math inside a program that is directed to signal processing. It means that any task when reimplemented with run-of-the-mill components + some software that improves ( some ) performance characteristic becomes patentable. > Clause (h) states “Method claims, whether independent or dependent, > reciting computer programs without process limitations in the form > of hardware features are not allowable. For a method reciting > computer programme to be patentable, it must clearly recite into it > limiting hardware integers that enable the program to function.” > > Clause (i) states “Claims directed at computer programs coupled to > hardware, enabling the hardware to perform a certain function may > be allowable, if such an invention meets all other conditions of > patentability.” Both the above are the same illogical contortions emanating from a gross misunderstanding of both programmable hardware and software. The fact is that ANY and ALL programmable hardware (including FPGA and other similiar bulk gate devices) are general puprose machines. Their use for some specific task is an artificial limitation born out of need for some performance characteristic, for example implementation of barrel shifters in a DSP with hardware caching of overflow bits. This allows certain mathematical functions to be implemented with great speed. Thus a program written for such a device is (by the PO definition) a software that is special and patentable. Now if someone uses the exact same software on a general purpose barrel shifter that does not cache the overflow bit and instead uses a program function to achieve caching, he will be violating the patent. Such a situation exists with GPU (or graphical processors), and SDR - Software Defined Radios -, both of which are a combo of hardware and software. It is entirely possible to use the code from such hardware (with suitable tweaks) on sufficiently fast general purpose hardware. Infact the SDRs initial implementations were on sound cards! > These clauses could be understood to mean that embedded software > can be patented. This could result in patenting of embedded > software where the improvement over prior art is only in the > software and not in the hardware part. An additional clause that “ > there should be improvement over prior art in the area of hardware > for an invention to be patentable” would have clarified the > position. Not at all. An embedded system is no different from any other programmable system when deconstructed as hardware and software. Embedded system are by design meant to be programmed. Consequently there should be no mention or claim whatsoever of the software for claiming inventiveness. As far as the patent office goes, it will strcictly consider such a combo to be a black box. Any subsequent claim of violation will have to be proved by the holder using his black box. Any attempt at deconstruction of hardware and software by the holder will automatically imply that substantial functionalty is present in the unpatentable software part and hence void the patent. Thus someone else using software and some other hardware in a non black box implementation will automatically be protected (he has declared his improvement to be substantially in software). The plus point in the above is that the patent office need not bother with trying to defend an undefendable postion, wholly artifically manufactured from fake logic. Applicants will actually start innovating, using software as a means of speeding up innovation, instead of playing sham games with paper, pencil and law. > > The clauses relating to Section 3(k) as extracted from the manual > are given below: > > ------------------------------- > > * Section 3(k) - **A mathematical or business method or a computer > programme per se or algorithms are not patentable.* > > a. Under this provision, mathematical methods, business methods, > computer programmes per se and algorithms are not considered as > patentable inventions. In relation to computer programs, the law > provides a qualification that what is not patentable is only > computer program per se. > > b. ‘*Mathematical methods*’ are considered to be acts of mental > faculty. A method of calculation, formulation of equations, finding > square roots, cube roots and all other methods directlyor > indirectly involving mathematical methods are therefore not > patentable. With the development in computer technology, these > mathematical methods are used for writing algorithms and computer > programs for different applications and the claimed invention is > sometimes camouflaged asone relating to the technological > development rather than the mathematical method itself. These > methods, claimed in any form, if in substance relate to > mathematical method are considered to be not 99 patentable. > > c. “*Business Methods*” claimed in any form are not patentable > subject matter.The term ‘Business Methods’ involves whole gamut of > activities in a commercial or industrialenterprise relating to > transaction of goods or services. With the development of Internet > Technologies, business activities have grown tremendously through > e-commerce and relatedB2B and B2C business. The claims are at times > drafted not directly as business method but apparently with > hitherto available technical features such as internet, networks, > satellites,telecommunications, etc. The exclusions are carved out > for all business methods and, therefore, if in substance the claims > relate to business methods, even with the help of technology, they > arenot considered to be patentable. > > d. Claims directed at ‘*computer programme products*’ are computer > programme per se stored in a computer readable medium and as such > are not allowable. > > e. If a claim in a patent application is not directed at a > computer programme per se it could be patentable, if all other > patentability conditions are met. This provision thus necessitates > distinguishing computer programmes per se from other types of > inventions that uses or implements computer programmes. Rubbish > f. The computer programmes are often claimed in the form of > algorithms as method claims or system claims with some ‘means’ > indicating the function of flow charts or process steps. The > algorithm related claims may be even wider than the computer > programme claimed by itself, for a programme represents a > particular set, the algorithm expresses the principles generally > and gives way for different programmes to be written based on the > same algorithm and as such are not patentable. If whoever wrote this, wrote the previous para, he can be accused of duplicity. > g. Essentially, combination all computer with some programmes > hardware need for a their functionality. In an application for > patent for a new hardware system, the possibility of a computer > programme forming part of the claims cannot be ruled out. It has to > be carefully considered as to how integrated is the novel hardware > with the computer programme. Further, it is also to be considered > whether the machine is programme specific or the programme is > machine specific. *A computer programme which may work on any > general purpose known computer does not meet the requirement of > patent ability*. The reason for special hardware are purely utilitarian. The software is required to use a "special" programmable device so that the combo device can run at lesser clock speed (consume lesser power, cheaper etc), to produce the technical effect. The same program with minor modification (in the case of well written programmes - more later) can run on a faster, (more power hungry, more expensive, larger memory etc) machine. Ironically a badly written program will be un rewriteable and will require substantial modifications to be portable. The PO is indulging in wishfull thinking that it can be an arbitrator of good / bad programming practices v/s inventiveness, while simultaneosly tripping itself with inconsistent definitions. Bad programming: You have such a situation existing, wherein a thoroughly rotten piece of software known as Microsoft Windows, has spawned an industry with very many patent claims on antivirus protection programs!! > > h. Method claims, whether independent or dependent, reciting > computer programs without process limitations in the form of > hardware features are not allowable. For a method reciting computer > programme to be patentable, it must clearly recite into it limiting > hardware integers that enable the program to function. > > i. Claims directed at computer programs coupled to hardware, > enabling the hardware to perform a certain function may be > allowable, if such an invention meets all other conditions of > patentability. By definition the purpose of a computer is to allow software to direct it to do something, which when coupled to additional devices known as peripherals produces physical effects like sound, light, movement etc. I have yet to see a program doing something when not coupled to hardware. Unless the PO considers steadily deteriorating storage media to be the real purpose of software and the "coupling to hardware to cause" as a marvelous new invention. These guys are tripping over themselves with definitions that are far removed from realty in order to somehow get software patents. The realty is ALL software including specifically those used in certain restricted types of hardware known as embedded systems become seperate entities when deconstructed in an attempt to prove innovativeness. Even hardware known as programmable logic exudes the exact same characteristics. The PO postion on patenting software is illogical, unjustified and untenable. -- Rgds JTD _______________________________________________ network mailing list [email protected] http://lists.fosscom.in/listinfo.cgi/network-fosscom.in
