Slashdot Mirror


Forgent and Microsoft Sue Each Other Over JPEG

goombah99 writes "CNET reports that the long running Forgent JPEG patent claim story has a new turn. Forgent Networks has filed a lawsuit against Microsoft, alleging the software giant infringed on its digital-image compression patent that serves as the technology behind JPEG. The suit comes in response to a suit Microsoft filed last week, asking the courts to find Forgent's patent unenforceable. '... despite Microsoft's recent inquiries about licensing the patent, they chose to file a lawsuit, leaving us no alternative but to assert infringement claims against it,' stated Richard Snyder, chief executive of Forgent. U.S. patent No. 4,698,672, relates to video image compression and transmission specifically and compression in general. The underlying technology is an amalgam of Cosine Transforms, Huffman coding, and odd details. Major corporations are respecting Forgent's claims: to date Forgent has collected about 100 million dollars in payments from computer and camera companies for this patent settling on suits with 31 companies. Past slashdot stories here, here and here. How might this impact Longhorn? Forgent has shown interest in selling it (to Compaq) so it's not unthinkable Microsoft could just buy it and own it."

18 of 296 comments (clear)

  1. This looks like a SCO tactic by Anonymous Coward · · Score: 5, Informative

    I think Forgent is playing a SCO tactic here; I have looked at the original patent and conclude that it is not patenting JPEG. In more detail, anything that JPEG uses which is detailed in this patent was discovered by somone besides the "inventor" of this patent.

    Forgent doesn't have a leg to stand on; as soon as this case hits the court room, this bogus patent will be declared invalid. The JPEG's group response to this nonsense patent.

  2. The Patent Expired in 2004 by GrassyKnowl · · Score: 2, Informative

    The patent was filed in 1987.

    A patent lasts 17 years so the patent is now expired.

    1. Re:The Patent Expired in 2004 by Thanatopsis · · Score: 2, Informative

      Patents usually expire after 20 years - Not 17. It also may have gotten extended. Patent expiration has no impact on past violations.

  3. Re:Why? by Anonymous Coward · · Score: 5, Informative

    This is not like the GIF situtation. What happened with GIF is that the lzv inventor patented his own invention. With JPEG, the patent Forgent has has nothing to do with JPEG, per se, except for the fact that the patent uses technology similiar to JPEG (DCTs, etc), all of which is technology invented before this particular Forgent patent.

    JPEG is an open format, plain and simple. Everyone who invented it made it an open format.

    Forgent does have a patent, but it is not for JPEG. It's for something else, and they're just playing the lawyer game to milk as much money from this patent as possible. It's too bad some companies caved in, giving these crooks money to hire more lawyers.

  4. Re:No, JPEG serves a purpose by Anonymous Coward · · Score: 1, Informative

    It says the primary purpose is to enable use of multiple JPEGs in an MNG which to me implies that it isn't going to be much different patent-wise.

  5. Re:Just when does this patent expire? by Husgaard · · Score: 2, Informative

    The US of A signed TRIPS, so they have to let patents be valid for at least 20 years according to article 33.

  6. Re:Just when does this patent expire? by Trepalium · · Score: 5, Informative

    The patent may have a life of up to 20 years. The rules to determine if the patent qualifies for a 20 year life or a 17 year life based on a filing from 1987 are too complex for me to figure out. All new patents filed in any TRIPS signatory nations have 20 year terms.

    --
    I used up all my sick days, so I'm calling in dead.
  7. Re:Why? by ortholattice · · Score: 1, Informative

    No, JPEGs can't be replaced by PNGs because PNGs (and GIFs) are lossless, whereas JPEGs are purposely lossy in order to achieve much better typical compression than PNGs (or GIFs). What you want to replace JPEGs with is the open-source, patent-free DjVu.

  8. Re:Why? by Anonymous Coward · · Score: 5, Informative
    Here's some more information, which I got from The web archive (It's not Karma whoring when your Mr. AC):
    Content of the Patent US 4698672 Relevant to JPEG

    The patent refers mostly to video compression. Some claims can be applied to still image compression as well. The central claims are formulated in Claim 1, 13, 38 and 39.

    These describe a well adapted code book used in the encoding process (Claims 1 and 13) and the decoding of the codestream (Claims 38 and 39), followed by entropy encoding. The key point of this algorithm is the use of so called "runs", sequences of identical symbols, encoded by the number and the value of the symbols in the run.

    Does JPEG use methods claimed by US-Patent 4698672?

    Despite differences in nomenclature and small deviations of the described algorithms the answer has to be in the affirmative, as both algorithms are close enough in details and in general. The similarities are as follows:

    1. The data stream is segmentated into runs of the most common symbol.
    2. Each run is characterized by a pair of symbols (n, a) with n indicating the length of the run and a information about the next symbol delimiting the run. This extends all the way to the last symbol.
    3. The remaining information is encoded using a special codebook.
    4. Runs of the most common symbol that reach all the way to the end of the block are encoded using the special EOB symbol.
    5. The codestream rewritten using the codebook is finally Huffman entropy encoded.
    The algorithm used in JPEG is very close to the technology described in Claims 1,13, 38 and 39. The two methods of encoding are not identical but similar to the point of algorithmic equality.

    Can "prior art" be proven in the case of Patent?

    The main ideas, methods and numerical experiments of the key parts of the invention, especially Claim 13 concerning the codebook content, had already been published in several articles before the registration of the patent.:

    1. H. Meyr, H.G. Rosdolsky and T. Huang, "Optimum Run length Codes, IEEE Transactions on Communications, vol.22, no 6, June 1974 , pp 826-835
    2. R.B.Arps, "The Statistical Dependence of Run-Lengths in Printed Matter", Nachrichtentechnische Fachberichte, vol 40, pp 218-226, 1971
    3. A.N.Netravali, F.W.Mounts and E.G.Bowen, "Ordering Techniques for Coding of Two-Tone Facsimile Pictures", The Bell System Technical Journal, Vol.55, pp 1539-1552, 1976
    4. H.Gharavi, Conditional Variable-Length Coding for Gray-Level Pictures, AT&T Bell Lab.Technical Journal, Vol. 63, pp 249 -260, 1984
    The general principle of encoding more common symbols in code of shorter length, and less common symbols in code of increased length has been in use since at least 1828 when Samuel Morse invented his alphabet. The Morsealphabet assigns longer code to less common symbols and shorter to more common ones, thus assign the shortest to e, then n, r, i, s, t. E Shannon formulated a mathematical theory of data compression using statistical correlations between spatially subsequent symbols and the encoding into code of length according to frequency 50 years ago.

    Does JPEG Infringe Upon US-Patent 4698672?

    First of all, despite the fact that the two algorithms are not truly identical they are similar enough to support the idea that JPEG does touch upon the claims made in the so called "Forgent-patent".

    But it can be safely assumed that JPEG does not infringe on the patent as "prior art" can be proven.

    The quoted literature does not describe the claims identically. But it does prove that even though the claims are differing in minor details from the ideas, methiods and results expressed within the articles the invention would have been obvious to an expert familiar with the ideas of the articles. They furthermore prove that the whole content of the invention was known to US scientists at the time of the registration of the patent.

    This should result in a rejection of any patent related claims by Forgent Networks.

  9. Editors Edited out key item in the post by goombah99 · · Score: 5, Informative
    The editors changed my submission omitting key details. The pattent issue is not cut and dried. Nor is neccessarily absurd patetn abuse. The pattent has two inventors. The first inventor, Dr Chen, is the developer of the Fast Cosine Transform and its applications to compression. Thus the fact that this pattent is based on the Fast Cosine transform is proof that prior art did not exist. The second inventor was also on the JPEG commission that created the standard, so it's reasonableto guess this techinology was put in the standard. The trouble is Dr. Chen published the compression algorithm a year before the patent was filed. In the US, the rule of thumb is you have a year to patent an idea after disclosing it, but exceptions can be granted.

    In this case the patent is mainly about compression and transmission of video. So its the combination of things not just the Cosine Transform.

    However a close reading of the patent claims shows that it also contains technology useful for still compression like jpeg. The fact that this is buried in the complex details may account for why it took so long for the assignee of the patent to get around to claiming it.

    The interesting thing here is that Forgent is willing to sel the patent. It already tried to sell the patent to Compaq. So in theory microsoft could pull a coup here not by breaking the patent but by purchasing it and thereby owning Jpeg. The fact that it already produces a revenue stream sort of underlies that. I also know that I the mpeg compression algorithm may also be based on Fast Cosine Transforms. THere is weaker chance it could be used to claim ownership of mpeg.

    --
    Some drink at the fountain of knowledge. Others just gargle.
    1. Re:Editors Edited out key item in the post by ergo98 · · Score: 2, Informative

      The interesting thing here is that Forgent is willing to sell the patent.

      Note that Forgent has already racked up $100 million in extortion, err licensing, fees relating to this patent, and they believe that they're just getting started. I doubt they plan on selling the patent for less than a billion dollars.

      Forgent is a classic patent enforcer - they have another patent related to PVRs, and they plan on using their JPEG warchest to finance that tax on consumer products.

      http://news.com.com/Patent+litigants+target+DVRs/2 100-1047-5659298.html

    2. Re:Editors Edited out key item in the post by rifftide · · Score: 2, Informative

      The text of patent 4,698,672 by W. Chen et al. (Compression Labs) makes reference to an earlier patent 4,302,775 by R. Widergren, W. Chen et. al. (Compression Labs). Patent 775 covers a method involving the application of the discrete cosine transform to blocks of an image or video frame. It was first filed in 1978 and apparently revised in 1981. I don't know whether that was the original DCT patent, but in any case it has expired. That's why Forgent hasn't claimed ownership of the basic DCT algorithm itself (which is fortunate, because that it the basis of practically all videoconferencing and streaming video systems in use today).

  10. Re:No, JPEG serves a purpose by Anonymous Coward · · Score: 5, Informative

    What a pathetic set of definitions.
    GIF. Lossless, but limited to 8 bit colour pallette.
    JPEG, can use lossy compression but true color.
    PNG, non-lossy, can use various pallette sizes including true-colour with full alpha channel. Can use a variety of compression algorithms, many of which are superior to that used by GIF.
    MNG. Friggen Google it. Who cares if it ain't well known - another patent war would be just push this spec needs.
    Hell. This set of definitions ain't great either, but is more accurate than the gibberish you typed.

  11. Re:Use It or Lose It by Anonymous Coward · · Score: 1, Informative

    While it is true that trademarks must be actively defended, or risk being lost, a reason why Kleenex, Xerox and Coke are constantly defending theirs, Copyrights and Patents can be exercised at the holders whim. This reflects the different nature of the forms of intellectual property, since Trademarks are indefinite, but patents, and copyrights (in theory), are of limited durations. You may therefore choose only to exercise your patents against Microsoft, but you must defend your trademark against every mom & pop store, or risk losing it.

  12. Re:This shows .. by close_wait · · Score: 2, Informative
    Microsoft has been on the right side of quite a few patent issues recently and I cant remember the last time they were on the wrong side of one
    ER, like the FAT patent they tried to enforce on all flash card manufacturers a couple of years ago?
  13. Re:Just when does this patent expire? by lottameez · · Score: 4, Informative

    They used to be 17, but got extended (and grandfathered) to 20 when the US signed up for the GATT treaty.

    --
    Yeah? Well I think you're overrated too.
  14. Re:Just when does this patent expire? by angle_slam · · Score: 3, Informative
    The rules to determine if the patent qualifies for a 20 year life or a 17 year life based on a filing from 1987 are too complex for me to figure out.

    The basic rule actually pretty simple. The term is either 17 years from the issuance of the patent or 20 years from the filing date. Calculate both dates. Whichever is later is the patent term.

    For patents filed after June 8, 1995, it's even simpler--20 years from the filing date. In the present case, the 20 year date is longer, so it will not expire until October 27, 2006.

    The more complicated part is determining if the maintenance fees are paid and if the patent in question is a continuation.

  15. Re:Why? by Anonymous Coward · · Score: 1, Informative

    How can they milk "decades old" technology when patents expire within 20 years?

    Patent holders only get about 16 years of temporary monopoly in exchange for sharing full, detailed, and *ENABLING* description of their invention with the public (via patent publication if granted) because the patent approval process takes about 3-5 years.

    Please understand that the goal of the patent system is to provide SUFFICIENT INCENTIVE to inventors to take substantial financial risks & personal sacrifices to solve tough problems and to fully disclose their solution with the public in exchange for a TEMPORARY (not decades long) monopoly.

    Like anything else, this noble goal which is often effective, is exploited by scoundrels who try to abuse the system and lack of time patent examiners have to do their work.

    ---
    Like patents, eliminating bugs in software makes more sense than abolishing all software in general.