RE: MS Windows and Unicode 4.0 ?

From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Tue Dec 02 2003 - 18:32:17 EST

  • Next message: Michael Everson: "Re: MS Windows and Unicode 4.0 ?"

    "Language Analysis Systems, Inc." writes
    > I'm not sure whether we're witnessing a customer-unfriendly change in
    > the business model (hardly the only one the computer industry has
    > foisted on us) or merely a transitional period while technologies are
    > still being developed. For my part, I'd like to see SVG fill this gap.
    >
    > The point is that while it shouldn't be Microsoft's job to solve
    > everybody's problems, it also shouldn't be the average Joe's problem to
    > solve them all himself.

    I agree with most of what you say, except that we are not exposing the
    problem faced by readers (yes they could buy a license to use thes fonts,
    but they will always think it should be for the documents they create
    themselves, not for the documents created by others they receive).

    So the main problem is for authors: they are not satisfied on relying on
    customers to have bought a separate license to read the document they sell:
    this limits the business of authors, in a way that did not exist the same
    authors chose to publish printed versions of their documents.

    So there's an urgent need for authors to get solutions. It's up to font
    designers to provide a texhnical solution that allows them to license their
    font designs to authors once, and give these authors the freedom to sell or
    broadcast the documents they create with them.

    That's why I think that font design providers (Adobe, Agfa MonoType, ...)
    should agree on a common format to allow authors to distribute freely the
    documents they create with these font designs. Then it's up to them to
    cooperate with operating system vendors so that these OS will be able to
    embed licensed fonts in documents, by managing two font stores in the OS:
    one for reusable licensed fonts that can be used and selected to create
    documents, one that get used as embeddable fonts (containing a signed copy
    of the license) when these documents are exported for publishing: the
    embedded fonts in the document then contain a signature of the license used
    by authors, and this signature cannot be removed from embedded fonts, nor
    detached from the document to create another document, without making the
    embedded font undecypherable/unusable.

    One way to achieve this is to only allow embedding of embeddable fonts
    within unmodifiable documents. This means a "export for publication"
    function in word processors, which should be the only way to create first a
    unmodifiable and signed document content, in which embedded fonts will be
    imported using the unmodifiable document content signature to encrypt the
    embedded font which will be attached to the document.

    The way these embeddable encrypted fonts are effectively attached to the
    signed document needs some other standard, but any composite document model
    could be used (including the standard RFC822 MIME multipart format used in
    Emails or in the .MHT file format for web archives).

    The document will still be modifiable, but this will require dropping the
    document content signature as well as the encrypted embedded font. Users
    that receive these documents and have a license for the referenced fonts can
    then work on it, and sign it agains with their own license. If the
    referenced fonts have no local license, it will be impossible to reuse the
    embedded tont design, and the user modifying the document will have to use
    another font and possibly alter the document style to work on it. But the
    same user will still be able to read it with the original design if the
    document is not altered.

    The difficulty is that each document will need its unalterable signature,
    and will need to embed a separate copy of the encrypted embedded font (this
    embedded font is not usable across multiple documents from the same author).
    The alternative on this could be that the documents could use a signature
    exposing the identity certificate of the author, and the OS could allow a
    encrypted embedded font to be used on all documents from that author by
    deciphering the document content with the key contained in a embedded font
    distributed separately.

    But this last solution would leave a security hole if a document can be
    created that matches the key of a separately embedded font: if it is not
    cryptographically strongly difficult to create such a document whose content
    would match the font key, the encrypted font could be reused. So the way to
    avoid it is to use very strong and quite long font keys. As the document
    will still need to reference that font key, the author of that modified or
    new document will need to steal an identity of another author to sign his
    modified document and expose that signature. This creates a vicious loop.
    That's why I think there's no way to embed fonts loaded separately from the
    document, and each published document wishing to use a licensed font design
    will need to embed that font design completely with their own encrypted
    version of that font, even if its the same one as in other documents from
    the same author.

    So encrypted embedded fonts should better have the force of a author
    signature, for the content using that embedded font. The best way to limit
    issues is to allow these encrypted embedded fonts to be viewed directly with
    a OS tool along with the clear-text details of the author contained in the
    signature certificate, in a way that says: "this signed embedded font is a
    version of the ABC-DEF font created by Typographer ABC is usable by
    documents created by XYZ, who has a valid license for using this font in the
    document authored by XYZ; this means that any document referencing this
    signed embedded font must be considered to be signed by XYZ and all the
    intellectual rights attached to the documents are specified by XYZ; if you
    think that the document you receive was not created by XYZ, you can first
    check <here> if the document structure and content is valid according to
    XYZ's control. If, not, then the distribution of this document to you is
    illegal, as it is both a breach in the font license by the real author of
    documents using this stolen font, or it may be an identity theft; if you
    think this is not the case, then you must first ask to XYZ to provide you
    the unencrypted font if he has the right to do so, or you must destroy the
    document using it unless you get a valid license for the font design ABC-DEF
    created by typographer ABC."

    __________________________________________________________________
    << ella for Spam Control >> has removed Spam messages and set aside
    Newsletters for me
    You can use it too - and it's FREE! http://www.ellaforspam.com





    This archive was generated by hypermail 2.1.5 : Tue Dec 02 2003 - 19:25:14 EST