[gnu.misc.discuss] The chilling effect of software patents

ghost@aladdin.com (L. Peter Deutsch) (01/30/91)

As some of you probably know, Adobe's Level 2 PostScript
specification includes facilities for data compression and
decompression according to the LZW (Lempel-Ziv-Welch) method.  I
would like to provide these facilities in Ghostscript.
Unfortunately, this method is claimed to be the subject of two
different patents, one by Unisys (which Adobe mentions in the new Red
& White Book, and has licensed) and one by IBM (which is claimed to
predate the Unisys patent by a week, but which is not mentioned in
the R & W Book).  Because of this, I am being forced to waste
otherwise productive time digging up addresses of patent departments,
writing to them, trying to figure out for myself which patent
applies, etc.  Furthermore, I may be required to pay a substantial
fee to license these questionable patents.  I am told, for example,
that the authors of PK-ZIP, one of the most popular PC compression
utilities, had to pay a fee of $10,000 for use of the LZW method.
This is completely out of the question for me, since I'm not making
any money from Ghostscript: if the fee is more than $100 or so, or if
the licensing terms preclude free distribution of Ghostscript,
Ghostscript will simply not include LZW coding, but will include some
freely available compression method (such as the Y coding method
posted to gnu.misc.discuss recently) instead.

Adobe's documentation also says that all ISVs who want to use LZW
compression may be required to license it from Unisys (or IBM,
depending on who you believe).  This may mean you, if you are writing
software that takes advantage of these facilities in Level 2
PostScript.  (I don't know the law about how far a piece of software
can get distributed before you have to license patents that may
apply: in theory, I guess you have to license even if no one but you
ever sees it.)

This is only the tip of a very large iceberg for the software
industry.  I hope you will think carefully before writing programs
that use patented facilities, because the costs to you and to others
may be much higher than you might think.

If you agree with me that software patents have a net harmful effect
on the industry, there are things you can do, and now is the time to
do them.  I don't think comp.lang.postscript is the place to continue
this discussion, so follow up to gnu.misc.discuss.

L. Peter Deutsch :: Aladdin Enterprises :: P.O. box 60264, Palo Alto, CA 94306
ghost@aladdin.com ; {uunet,sun,decwrl}!parcplace!aladdin!ghost ; (415)329-0264
	   >>>>>> New phone # after 2/3/91: (415)322-0103 <<<<<<
	    "Implementation is the sincerest form of flattery."

jbuck@galileo.berkeley.edu (Joe Buck) (02/09/91)

This started in gnu.misc.discuss.

This is also going to comp.org.usenix; you'll see why below.  If your
followup is appropriate to only one group, please edit the Newsgroups
line.

In article <3735:Feb822:24:5991@kramden.acf.nyu.edu>, brnstnd@kramden.acf.nyu.edu (Dan Bernstein) writes:
|> In article <b0hp2o.e8d@wang.com> lee@wang.com (Lee Story) writes:
|> > (1) Challenge the patents or copyrights.  This probably should be open
|> >     to all, not just those with the deepest pockets, but is it?
|> 
|> There are two basic ways to challenge a patent.
|> 
|> 1. Send $2000 to the Office of Patents and Trademarks with a challenge.
|> The challenge typically goes through each claim of the patent and points
|> out where that claim appeared in the prior art. Occasionally a claim may
|> be declared invalid for other reasons (such as ``it describes a
|> mathematical algorithm,'' not that I know of this ever succeeding).

I hadn't known about way #1.  I was just thinking this morning about the
problems certain software patents pose and the need for overturning
some of them, and decided that we (the software communitity) need
something in addition to the LPF approach of lobbying against software
patents and user interface copyrights.  Even if this succeeds, we need
to overturn certain patents already on the books.  Perhaps a foundation
can be founded to fight patents that are generally recognized to be
evil.  Example #1, the recently granted include-file patent, which
puts every C, C++, and Fortran compiler in violation.

Possibly Usenix can get into the act.  I have a mailing from them in
front of me that says the following (as part of a membership survey):

Software Patents

The US Patent Office has recently issued patents on some disturbing things --
like the concept of an include file.  Originally they ruled that a mathematical
theorem could not be patented.  In the transition from theorem to algorithm
to implementation, the ruling was reversed.  The patent on an include file,
applied for over 15 years ago, was just recently issued.

11.  Please check all the roles, if any, that you feel USENIX should play in
software patent issues.

[ They list four alternatives ]

	Pro-active: help contest software patent applications
	Informative: produce patent "snitch" reports
	No software patent activities
	Do not care.

I checked "pro-active".  If Usenix doesn't want to take this job on, it
seems to me that it's time to form a foundation to collect the $2000,
get a good writeup on the prior art on include files which goes back
a lot longer than 15 years, and blow this idiot patent away.

Then we can take on Hayes' "+++" patent.  Seems those guys have gotten
rid of their good engineers (they recently hired a consultant to teach
Hayes employees how modems are designed, and I'm not kidding) and decided
to hire good lawyers instead.

--
Joe Buck
jbuck@galileo.berkeley.edu	 {uunet,ucbvax}!galileo.berkeley.edu!jbuck	

brnstnd@kramden.acf.nyu.edu (Dan Bernstein) (02/09/91)

In article <10957@pasteur.Berkeley.EDU> jbuck@galileo.berkeley.edu (Joe Buck) writes:
  [ quoting a Usenix mailing ]
> Originally they ruled that a mathematical
> theorem could not be patented.  In the transition from theorem to algorithm
> to implementation, the ruling was reversed.

Uh, no. To my knowledge the original precedents are still going strong.
The problem is that the Patent Office folks don't seem to be able to
recognize mathematics when it's shoved in their face.

> I checked "pro-active".  If Usenix doesn't want to take this job on, it
> seems to me that it's time to form a foundation to collect the $2000,
> get a good writeup on the prior art on include files which goes back
> a lot longer than 15 years, and blow this idiot patent away.

On the other hand, once you have a good enough writeup, you can just
distribute that and save the $2000. Something based on the prior art
doesn't infringe a patent: if you get sued, you just show the court the
references to prior art and they won't even take a second glance at the
patent.

It would be much more interesting to challenge the compression patents,
on several grounds: they describe purely mathematical inventions; they
claim all implementations of the methods; and, of course, they are
obvious from the prior art. To convince a court of these things would
take a lot of work but would be well worth it.

The best solution would be for the LPF to push through legislation
eliminating these patents permanently.

---Dan

henry@zoo.toronto.edu (Henry Spencer) (02/10/91)

In article <7802:Feb910:14:3891@kramden.acf.nyu.edu> brnstnd@kramden.acf.nyu.edu (Dan Bernstein) writes:
>The best solution would be for the LPF to push through legislation
>eliminating these patents permanently.

I'd rather see a shortening of patent length and a loosening of terms than
a complete ban.  The alternative to patents, all too often, is secrecy.
-- 
"Maybe we should tell the truth?"      | Henry Spencer at U of Toronto Zoology
"Surely we aren't that desperate yet." |  henry@zoo.toronto.edu   utzoo!henry

smb@ulysses.att.com (Steven Bellovin) (02/16/91)

In article <10957@pasteur.Berkeley.EDU>, jbuck@galileo.berkeley.edu (Joe Buck) writes:
} Possibly Usenix can get into the act.  I have a mailing from them in
} front of me that says the following (as part of a membership survey):
} 
} Software Patents
} 
} The US Patent Office has recently issued patents on some disturbing things --
} like the concept of an include file.

Speaking of which -- does anyone have the patent number for this patent?
I'd like to see for myself what it really says.

waters@nddsun1.sps.mot.com (Mike Waters) (02/16/91)

In article <14323@ulysses.att.com> smb@ulysses.att.com (Steven Bellovin) writes:
>In article <10957@pasteur.Berkeley.EDU>, jbuck@galileo.berkeley.edu (Joe Buck) writes:
>} Possibly Usenix can get into the act.  I have a mailing from them in
>} front of me that says the following (as part of a membership survey):
>} 
>} Software Patents
>} 
>} The US Patent Office has recently issued patents on some disturbing things --
>} like the concept of an include file.
>
>Speaking of which -- does anyone have the patent number for this patent?
>I'd like to see for myself what it really says.

How about a synopsis, title or even some keywords for a search?
A date of issue would even narrow it down.

Typically these patents are issued with a title that doesn't give a lot
of clues about the terms coined (later) which describe their use.

For example the (in)famous "spreadsheet patent" #4,398,249, issued Aug. 12
1970 is titled "Process and Apparatus for Converting a Source Program
into an Object Program". Basically a compiler for a program written as
interacting cells - we now call it a spreadsheet.

I would be willing to do a reasonable amount of work to track this down,
but not to search almost a million patents by hand!

-- 
           *Mike Waters    AA4MW/7  waters@nddsun1.sps.mot.com *
Chemicals, n.:
	Noxious substances from which modern foods are made.

jim@melmac.umd.edu (Elie W. Mansour) (02/16/91)

In article <14323@ulysses.att.com> smb@ulysses.att.com (Steven Bellovin) writes:
>
>Speaking of which -- does anyone have the patent number for this patent?
>I'd like to see for myself what it really says.

I can get you the full text of the patent if you or someone can give me names
of any of the parties, or any other more tangible info.

Thanks,

Elie
	
-- 

---|-----------------|---	----------------------------------
	Elie Mansour		| 				|
George Washington University	| Internet:jim@melmac.umd.edu	|
    National Law Center		| Bitnet:chandler@gwuvm		|
     (202) 994-4943		|				|
---|-----------------|---	---------------------------------