IFX Mail Archive: IFX> IP issues around Filters in PDF.

IFX> IP issues around Filters in PDF.

From: Rick Seeler (rseeler@adobe.com)
Date: Wed Sep 18 2002 - 13:25:57 EDT

  • Next message: Rick Seeler: "IFX> PDFax Image format 'tree'."

    Here's a question I posed to some of our engineers who deal with IP
    issues around PDF:
    -------

    My Question:
    ...
    I need to find out what kind of IP coverage creators and/or readers of
    PDF have with regard to some of the PDF filters. The filters I've
    listed below implement image compression algorithms that usually have IP
    issues around them. Does the fact that someone is following the PDF
    reference in implementing PDF give that entity any rights to use those
    algorithms? If not, what kind of licensing or IP issues would they be
    confronting? Is there anything already written up about this?

    - FlateDecode
    - JBIG2Decode
    - DCTDecode (JPEG)

    We're trying to choose minimums for PDFax and need to pick royalty free
    requirements. Any help or direction is appreciated.

    ...and the responses...
    > ...
    > The reason that the PDF specification mentions the Unisys LZW
    > patent is that we describe the actual algorithm for the LZW
    > method, part of which is subject to the patent claims. Thus,
    > it is necessary to warn readers about this.
    >
    > The filter algorithms you listed (Flate, DCT, JBIG2) are all
    > described in
    > external specifications. Any relevant IP matters ought to be
    > described there; they are not mentioned in the PDF specification.
    >
    > I can say with reasonable confidence that FlateDecode is free
    > of any IP encumbrance. That was the whole point of creating
    > this filter. It is the same as the public-domain zlib/deflate
    > filter, which was carefully designed to avoid any known
    > patents. Of course, a submarine patent could still emerge,
    > but I can't judge the likelihood of that.
    >
    > There are said to be patents covering certain aspects of DCTDecode
    > (JPEG) and JBIG2Decode; however, I can't offer any specific
    > information about this.
    >
    > Even when the compression method itself is free of IP
    > encumbrance, there might nevertheless be patents on specific
    > techniques used to implement the method.
    > ...

    AND

    > ...
    > I think your question is very similar to one that we used to get with
    > PDF 1.0-1.1 regarding use of LZW. At that time, Unisys held the patent
    > (they still do) and although LZW was specified in PDF, any developer
    > of software to read/write/manipulate PDF had to obtain an IP license
    > from Unisys. Many of our early developers had trouble doing so (either
    > due to terms, or pricing, or both) and that motivated us to put Flate
    > into PDF so that it would be easier for developers. We try to use
    published
    > standards, but yes, some of them have IP issues (not all standards
    committees
    > are successful at the royalty-free IP thing.)

    > I think the committee (PWG) will have to investigate any potential IP
    issues
    > with each of the compression algorithms you've identified. Adobe
    doesn't
    > provide that kind of guidance.
    > ...

    ------------------

    Also, I will be getting a statement from Adobe legal that is created for
    the PWG, specifically, that should address any outstanding concerns
    about IP around Adobe's PDF patents. I've seen a draft of this document
    and it looks pretty thorough. I will send an e-mail with the statement
    as soon as it has been approved.

    -Rick



    This archive was generated by hypermail 2b29 : Wed Sep 18 2002 - 13:26:07 EDT