This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Decompiler Project and Mailing List

   My apologies to the lists for an unwanted thread.  I don't know which
list Joe is reading from, or I'd winnow the cc: list.  Unfortunately, I
can't let this pass without comment.

On Wed, 17 Nov 1999, Joe Buck wrote:
> While a decompiler is certainly a useful tool, producing copylefted
> streaming tools by reverse-engineering proprietary codecs, as the
> website advocates, is the wrong approach.  The problem
> is patents; you can't use a patented technique without licensing the
> patent, or you put yourself at considerable financial risk.
> Furthermore, you can hardly later claim accidental infringement when the
> big boys come with their lawyers when you've made clear that you've
> obtained your tools by reverse engineering.
    You are correct that there will likely be a lawsuit.  However, I think
you would be assuming too much if you think a software patent holder will
necessarily when.  The argument is on 2 grounds: (1) there is a conflict
between patent law and the 1st amendment when it prevents me from
expressing ideas in computer languages (see the Bernstein case), and (2)
restraining free software (GNU sense, not zero-cost) in favor of patents
might not serve the public interest in the way that the constitutional
clause granting Congress the power to establish patents demands (i.e. to
advance the useful arts and sciences).  
     While the idea of a prolonged legal battle is not my idea of a good
time, the idea that I should give up my civil liberties to write (in
whatever language I want, and whatever type of ideas I want) and help
others with that writing just so some companies can turn a profit is
a lot worse.

> The only way to produce free streaming media is to use algorithms not
> covered by patents.  This is a difficult job, but far from impossible.
> Alternatively, all development could be done in countries that don't
> legally honor software patents, but the result would be tools that
> people in the US and a number of other countries cannot legally use.
     This will be done as well.  Actually, it's much more interesting than
the reverse engineering parts (well, writing a decompiler is a very
interesting problem itself), which I'm mainly doing because of the large
amount of content that people already have in these proprietary formats.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]