[odf-discuss] On covenants

[Alex Hudson]

marbux wrot:
>
>     Going back to real-world examples: the MP3 specification is for an
>     encoded data stream. It doesn't tell you how to decode or encode any
>     given piece of audio, and in fact there are a variety of encoders and
>     decoders available which work in completely different ways. They all
>     infringe the MP3 patents, though.
>
>
> I do not comprehend what this has to do with your argument that the
> patent  "is of necessity in the code."

The claim that the patent makes must be in the code for the patent to be
infringed. In the example of MP3, one of the things they claim is a
method of quantizing audio (losing data, basically) such that the result
is optimal to the human ear: you convert audio into the frequency
domain, and lose more data in those frequencies humans can't hear, more
or less.

The patent claim doesn't tell you _how_ to do that specifically, but any
specific code for doing that is covered by that patent claim, and there
is no way of doing it which doesn't infringe that claim. I would say
that patent claim is necessary to conform to the MP3 specification.

> Try it this way; there are nations where software patents are not
> recognized and thus there can be no infringement. Is the same MP3
> patent claim still "of necessity in the code" when the MP3
> encoder/decoder is executed on computers in such nations?

Yes, of course it is - the patent claim is just a method for achieving
something. Whether or not that method is elucidated in a valid patent is
irrelevant, the method is still in use by the software.

>     > Yup. Now tell me how you read the  word "infringes" into the CNS. It
>     > isn't there and would have to be in order for the sentence to
>     make sense.
>
>     Patent claims necessary to conform to the technical specification are
>     those patents that would necessarily be infringed by any
>     implementation
>     of the specification. 
>
>
> Oh, perhaps you might direct me to the definition in the CNS that says
> this is so?  You are implying a definition that is not there and that
> is expressly forbidden by the "no implications" sentence. 

This is not going to get us anywhere.

All I can do to explain to you how to understand that sentence is by
drawing analogy and expounding. If a sentence said "all things yellow
are allowed", it's like me saying "for example, red things are
disallowed" and you saying "it doesn't say anything about red things, so
that interpretation must be incorrect".

If all you allow me is the words in the CNS, then I can nothing to
further your understanding of it because you immediately dismiss any
other form of words, and merely repeating the CNS to you is unlikely to
spark any kind of eureka moment.

> Microsoft's relevant "loophole" (there are others) is the fact the
> following string is an empty set, "any of its patent claims necessary
> to conform to the technical specifications for the Microsoft Office
> 2003 XML Reference Schemas." It is grammatically the equivalent of
> saying "any of its apples necessary to conform to the technical
> specifications for an orange." Both sets are empty. You can only get
> some other result by implying words that are not there and that is
> forbidden by the "no implications" sentence if to do so would create a
> right not expressly granted.  As written, Microsoft has conveyed no
> rights since no patents are necessary to conform to a file format
> specification.

If this is your interpretation, I disagree with it on two fronts:

    * I believe it is likely that patent rights are required to conform
      to OpenXML; I don't believe that I am alone in this either. It is
      also exceptionally difficult to show that no patents apply, since
      it's an attempt at proof of non-existance.
    * I don't have a problem with "patent claims necessary to conform".
      You seem to interpret this as saying "software which has the text
      of the patent claims written into it" - I interpret it as
      "software containing methods claimed by patents". I argue the
      former interpretation is unintelligible, not the latter.


>     On first reading, it would seem that it is, but then Sun goes on
>     to say
>     that their grant only covers OpenDocument implementations. You
>     couldn't
>     write a Java interpreter that natively read/wrote OpenDocument,
>     and use
>     that patent grant to protect yourself from Sun's patents on the Java
>     technology.
>
>
> Excuse me. How on Earth did you come to that conclusion? Isn't the
> Java interpreter still included in the universe of "any
> implementation" of the ODF specification? A little light on the
> subject, please.

I struggle to see how a Java interpreter could be defined as an "ODF
implementation". Note that Sun said "implementation", not "application"
or "software product" or any other phrase which means "the whole thing".

I do not read Sun's covenant as being a blanket grant to their entire
patent portfolio for any product which contains some ODF conformance.
They say "implementation of ODF" - to say that it covers those parts of
software which are not implementations of ODF is to imply words into the
grant which aren't there.

-- 
Alex Hudson
IT Director, Stratagia Ltd.
P: 0845 226 17 13     W: http://www.stratagia.co.uk/
F: 0845 226 17 14     E: alex (at) stratagia.co.uk