Patents Unleashed and the future of Java Programming

[---]

Hi guys,

Yesterday EU did put a tombstone onto the free and open source
programming in Europe, and i think open Java (and C#) as well.

From now on, or from the final vote on, large companies will find
a lot of earnings simply dissecting the code of open source and java
programmers, to extract patentable ideas and put them under
a patented protection.

How will we, Java developers, create large systems with a language
that can be easily decompiled to seek for patent breachings of
very simple ideas and algorithms?

I foresee a world of lawyers instead of inventors, since code
dissection to search implementation of algorithms can be made
authomatic for easy to decompile languages like java and Mono C#...


Damn

 

[ak]

Yesterday EU did put a tombstone onto the free and open source

programming in Europe, and i think open Java (and C#) as well.

From now on, or from the final vote on, large companies will find
a lot of earnings simply dissecting the code of open source and java
programmers, to extract patentable ideas and put them under
a patented protection.

How will we, Java developers, create large systems with a language
that can be easily decompiled to seek for patent breachings of
very simple ideas and algorithms?

I foresee a world of lawyers instead of inventors, since code
dissection to search implementation of algorithms can be made
authomatic for easy to decompile languages like java and Mono C#...

demokratie is good.., if you can buy parliament!

 

[Roedy Green]

The good news in this era should be temporary. Eventually the patents
on everything obvious should expire.

 

[Chris Smith]

The good news in this era should be temporary. Eventually the patents
on everything obvious should expire.

Unfortunately, no. It's surprising how often people come up with new
motivations to patent some obvious idea. I recently spoke to someone,
for example, whose company has patented the idea of using colored lines
between boxes to represent relationships in a diagram generated by their
software.

--
www.designacourse.com
The Easiest Way to Train Anyone... Anywhere.

Chris Smith - Lead Software Developer/Technical Trainer
MindIQ Corporation

 

[Chris Smith]

Do you mean that the EU has made it possible to patent ideas that are
already part of what the lawyers call "prior art"?
In the U.S., you can't patent something that can be proved to be
prior art. Or, if you manage to get something past the patent examiners,
someone else can get it overturned it they care about it.
An EU patent on prior art would be unenforceable elsewhere.

That's the theory. Unfortunately, not everyone has the money to hire
lawyers... The unfortunate truth about software patents is that:

1. Many companies don't pay attention to prior art, but instead apply
for patents on everything they possibly can.

2. Most patent organizations don't have the time or resources to get the
kind of expertise needed to determine whether an algorithm is really
novel or simply an obvious modification on some prior algorithm.

3. Because of this, even more companies frequently acquire frivolous
patent portfolios in order to fight back against possible legal attacks.

4. There are plentiful examples of patents that are clearly not novel,
but have yet to be successfully challenged in any court, simply because
it's so expensive to do so.

More details are plainly available from your average Google search.

--
www.designacourse.com
The Easiest Way to Train Anyone... Anywhere.

Chris Smith - Lead Software Developer/Technical Trainer
MindIQ Corporation

 

[Robert Metzger]

Do you mean that the EU has made it possible to patent ideas that are
already part of what the lawyers call "prior art"?
In the U.S., you can't patent something that can be proved to be
prior art. Or, if you manage to get something past the patent examiners,
someone else can get it overturned it they care about it.
An EU patent on prior art would be unenforceable elsewhere.

1) Encourage your friends to publish all their patentable ideas that they
have already used in open source.
2) Create new venues for the publication of those ideas.
3) Educate the EU patent examiners about prior art.
4) Run all your classes through the best available obfuscator (see below).

As the author of a book on algorithm recognition
("Automatic Algorithm Recognition and Replacement:
A New Approach to Program Optimization", Metzger & Wen, MIT Press, 2000)
I can tell you that algorithm recognition of bytecode is a **VERY** hard
problem.

Our work did algorithm recognition based on a high level intermediate
representation generated by an optimizing compiler. Not only did we
have all the original content from the original source, but the analysis
of an interprocedural optimizer as well. It was still a very hard problem,
and we limited the scope of the work in order to achieve success.

Byte code has lost significant semantic content compared to either a compiler
intermediate representation or original source code. This will make it
that much more difficult to recognize any algorithm worth patenting.

Obfuscators can be built that will make algorithm recognition of byte code
much more difficult than it is on unobfuscated bytecode.

 

[Roedy Green]

Unfortunately, no. It's surprising how often people come up with new
motivations to patent some obvious idea. I recently spoke to someone,
for example, whose company has patented the idea of using colored lines
between boxes to represent relationships in a diagram generated by their
software.

At some point there has to be political pressure to stop patent
offices from issuing these nuisance software patents.

I think Amazon patented one click purchasing, as if reducing the
number of clicks was not an obvious design goal.

By the way, one click installs are prior art.

 

[George Neuner]

In the U.S., you can't patent something that can be proved to be
prior art. Or, if you manage to get something past the patent examiners,
someone else can get it overturned it they care about it.

The first problem is that US examiners don't know where to look for
prior art. IEEE and ACM have been working for a number of years to
put together a prior art database for them but it is still very
incomplete.

The second problem is that US examiners are required to grant a patent
if they cannot find prior art. CS is still more of an art than a
science and much knowledge has been passed by tradition and never
published. The result is that a lot of techniques that everyone knows
and uses are now suddenly being patented by unscrupulous companies who
have wet dreams of getting a 5 cent royalty every time an array is
indexed.

AFA getting an issued patent overturned ... that is *VERY* difficult
and expensive to do. It is presumed that the patent holder already
adequately proved the case and the bar is set very high for anyone
seeking to disprove it. No one will do it who doesn't have a lot of
money invested in whatever the patent addresses.

As the author of a book on algorithm recognition
("Automatic Algorithm Recognition and Replacement:
A New Approach to Program Optimization", Metzger & Wen, MIT Press, 2000)
I can tell you that algorithm recognition of bytecode is a **VERY** hard
problem.

Good to know.

George
=============================================
Send real email to GNEUNER2 at COMCAST o NET

 

[Andrew Thompson]

CS is still more of an art than a
science and much knowledge has been passed by tradition

...like what?

..and never
published.

What exactly do you mean by 'published'?

Published in a peer reviewed Computer Science
monthly publication?

Published in a 'Teach yourself in 24Hrs' book?

Published in a web-page, or in a public forum?

 

[Roedy Green]

The second problem is that US examiners are required to grant a patent
if they cannot find prior art.

I met a guy who worked in the Swiss patent office. He said he has
four days to decide. The typical application is 1.5" thick.

Consider also the authors have done the best they could to baffle with
bullshit their simple idea making it sound amazing and novel.

Patents were originally supposed to encourage innovation. They are
having the opposite effect. Only a large company has enough of a
palette to interest others in cross-licensing.

Patent holders sit on ideas and play dog in a manger. Patent holders
stop people from using obvious ideas.

I think the public should be involved in patent application. When a
patent has received tentative approval, it is published
electronically, and the public has a month to respond. The patent
office makes the final decision. That way the community can help with
the prior art search.

 

[Sudsy]

Roedy Green wrote:

I think the public should be involved in patent application. When a
patent has received tentative approval, it is published
electronically, and the public has a month to respond. The patent
office makes the final decision. That way the community can help with
the prior art search.

For an idea that's so simple, it's brilliant! Talk about an "appropriate
use of technology". Anyone in a position to pitch this to the powers-
that-be? Certainly the agencies involved have enough computing power
for the task. Heck, if e-mail services are offering gigabyte mailboxes
then we know that storage costs are not onerous.
Roedy, with the writ expected to drop this Victoria Day weekend why
not run for MP?

 

[Jim]

I met a guy who worked in the Swiss patent office. He said he has
four days to decide. The typical application is 1.5" thick.

Consider also the authors have done the best they could to baffle with
bullshit their simple idea making it sound amazing and novel.

Patents were originally supposed to encourage innovation. They are
having the opposite effect. Only a large company has enough of a
palette to interest others in cross-licensing.

Patent holders sit on ideas and play dog in a manger. Patent holders
stop people from using obvious ideas.

I think the public should be involved in patent application. When a
patent has received tentative approval, it is published
electronically, and the public has a month to respond. The patent
office makes the final decision. That way the community can help with
the prior art search.

Here, Here. Especially in the case of software patents. I really doubt
that there are many "really original ideas" out in software land. How
often do you really see a really novel something like "Quicksort" pop
up?

Jim

 

[George Neuner]

..like what?

Have you ever seen documented an algorithm for determining the number
of '1' bits in a word? The highest '1' bit in a word? The lowest?

When you learned to program you a) didn't care, b) figured them out
for yourself or c) somebody showed you how and you never looked back.
In 25 years, I've never seen these algorithms covered in ANY book or
paper targeted for ANY level of programmer. Yet we all know how to do
these elementary things.

FYI, software solutions to the above are patented by NEC.

What exactly do you mean by 'published'?

I mean "published" as in disseminated in some (reasonably) public
forum. For this purpose I will accept internal company efforts meant
to educate employees but not necessarily distributed beyond the
company. However, I disqualify "reading coworkers code", "looking at
your coworkers notes", etc. and any other sources not meant to be
disseminated in a general way.

There is an enormous amount of "gestault" knowledge about what works
and what doesn't which we all figured out for ourselves by trial and
error, was demonstrated to us by some mentor or was grokked by reading
someone else's code. We in turn have showed others what works and
what doesn't. No doubt we learned by reading the published works of
others, but our work is unique - an amalgam of our own experience
which is seldom written down.

Pick any non trivial problem and take a moment to think about how you
would go about solving it. Ok, now you've though about it. So tell
me - where did you see that solution published?

Even the most anal retentive documentation efforts (I've worked on
medical diagnostic instruments) fall far short of documenting the
software algorithm used to determine whether an integer value is odd.

George
=============================================
Send real email to GNEUNER2 at COMCAST o NET

 

[George Neuner]

Roedy Green wrote:


For an idea that's so simple, it's brilliant! Talk about an "appropriate
use of technology". Anyone in a position to pitch this to the powers-
that-be? Certainly the agencies involved have enough computing power
for the task. Heck, if e-mail services are offering gigabyte mailboxes
then we know that storage costs are not onerous.
Roedy, with the writ expected to drop this Victoria Day weekend why
not run for MP?

It will never happen in the US. The "secret" aspect of application
was intended to foster innovation by preventing big companies from
monitoring pending patents from small inventors and opposing them for
sport. Where it backfires is in disciplines like software and biotech
which are constantly evolving and where there is no large database of
education and prior art for examiners to query - this creates a
knowledge vacuum where the right spin can get almost anything
patented.

There were some countries that used to have opposition periods - not
sure what has happened what with the EU taking countrol - but it was
expensive to oppose or even comment on a pending patent. The process
wasn't really "public" in that everyone involved had to sign NDAs to
protect the rights of the filer (who may have disclosed other
protected matters in the course of the application). Therefore very
few pending patents were ever opposed - most companies chose to wait
until the patent was public and then fight it openly.

George
=============================================
Send real email to GNEUNER2 at COMCAST o NET

 

[Andrew Thompson]

...
Have you ever seen documented an algorithm for determining the number
of '1' bits in a word? The highest '1' bit in a word? The lowest?

No. I did a quick search on Google groups and
my most specific searches failed to unearth
that exact algorithm.

My more general searches showed 1000+ hits, so
I did not pursue it further.

Which brings me to the specific question.
Do you consider usenet in, or out?
Should usenet be considered a form of publishing?

 

[Roedy Green]

Which brings me to the specific question.
Do you consider usenet in, or out?
Should usenet be considered a form of publishing?

It should be. It is far easier to search than printed works.

It documents the "obvious" stuff that newbies ask that never gets
published.

 

[Andrew Thompson]

It should be. It is far easier to search than printed works.

It documents the "obvious" stuff that newbies ask that never gets
published.

Which is where I am coming from. There would
be _very_ few things that were not both asked
and answered in public forums.

 

[Knute Johnson]

That's the theory. Unfortunately, not everyone has the money to hire
lawyers... The unfortunate truth about software patents is that:

1. Many companies don't pay attention to prior art, but instead apply
for patents on everything they possibly can.

2. Most patent organizations don't have the time or resources to get the
kind of expertise needed to determine whether an algorithm is really
novel or simply an obvious modification on some prior algorithm.

3. Because of this, even more companies frequently acquire frivolous
patent portfolios in order to fight back against possible legal attacks.

4. There are plentiful examples of patents that are clearly not novel,
but have yet to be successfully challenged in any court, simply because
it's so expensive to do so.

More details are plainly available from your average Google search.

Patents should never have been allowed for algorithms or code. That's
what copyright is for.

 

[George Neuner]

No. I did a quick search on Google groups and
my most specific searches failed to unearth
that exact algorithm.

My more general searches showed 1000+ hits, so
I did not pursue it further.

I don't have the cites for the patent numbers. They were granted
several years ago. I had some correspondence with various members of
ACM regarding them and the prior art database, but the mails are on an
old hard disk in my closet. If I can scare up an uninterupted hour or
two I'll see if I can fish them off.

Which brings me to the specific question.
Do you consider usenet in, or out?
Should usenet be considered a form of publishing?

I said previously that I would accept web pages - but now thinking
about it some more I would have to say not. My position on this
extends to Usenet as well.

The problem with Usenet is the sparsity of archiving. Many Usenet
groups are not archived at all and the ones that are have an epoch
beyond which searching is impossible. Only a few moderated groups are
fully archived and searchable in their entirety. The oldest archives
at uu.net go back only to the 80s (Usenet began in the 60s) and have
you ever tried to find something in there? Google's coverage is good
but uneven: the indexes are windowed by storage and are weighted
toward more popular groups, many groups are simply not indexed at all
and the earliest stuff available is circa 1989.

The Web is even worse because there is virtually no attempt being made
to preserve it for posterity. Links vanish, content is "refreshed" or
"updated", pages are reformatted, static content becomes dynamic, etc.
- all with amazing rapidity. It's been estimated that at any
particular time, 40-50% of the cited URLs are broken. And I've read
guesstimates that up to 90% of the total web is completely unindexed.

This dynamism greatly affects the ability to do research. One of the
more important precepts of publishing is that edition X doesn't
obliterate edition X-1 ... the previous edition is always there for
comparison. When something is written in a book, you have a pretty
good assurance that, barring the end of civilization, the book will
always be somehow available (at least for the last couple of centuries
and even if you have to physically visit the library that entombs it).
You have virtually no assurance for anything on the Internet unless
the particular content you reference happens to be in a deliberately
maintained library. Even then there is the question of whether any
privately maintained library will still exist umpteen years from now.

I would like to consider the Internet publishing - but until there is
some, probably government sponsored, effort to mirror and organize all
that content, I really think that, in general, it simply can't be
considered publishing.

George
=============================================
Send real email to GNEUNER2 at COMCAST o NET

 

[Joona I Palaste]

Have you ever seen documented an algorithm for determining the number
of '1' bits in a word? The highest '1' bit in a word? The lowest?

When you learned to program you a) didn't care, b) figured them out
for yourself or c) somebody showed you how and you never looked back.
In 25 years, I've never seen these algorithms covered in ANY book or
paper targeted for ANY level of programmer. Yet we all know how to do
these elementary things.

FYI, software solutions to the above are patented by NEC.

A really simple way to determine the highest '1' bit in a word is to
keep right-shifting it until it becomes 0. Similarly, a really simple
way to determine the lowest '1' bit is to keep right-shifting the word
until it becomes odd, or 0 (in which case there were no '1' bits).

Are you telling me that if I implemented those algorithms out in actual
code, I would be stealing NEC's intellectual property?

--
/-- Joona Palaste ([email protected]) ------------- Finland --------\
\-- http://www.helsinki.fi/~palaste --------------------- rules! --------/
"Parthenogenetic procreation in humans will result in the founding of a new
religion."
- John Nordberg