<< We’re starting to assemble; it seems like everyone’s Food for Thought Dinners went well; certainly we had to be "asked to leave" to make room for more clientele >>
"rights in software + the SCO case" — the title slide for “Z”
"Intellectual Property Protection for Software" — the title slide for Terry
And it looks like Terry is “lead off” speaker — first, announcements
this is a new room; how to work the mikes
we’re handing out a questionnaire; turn it in to get your certificate
The topic for the morning is software; I’m going to concentrate on the IP aspects; Z’s will concentrate on open source and the current SCO controversy
It will be interwoven, so look out!
Basic propositions:
1) What kinds of activities might a s/w developer want to control?
- consumer reproduction of object code
- commercial reproduction of object code
- incorporation of parts of source code in new programs
- preparation and distribution of improved versions/derivative works
Note that these are going to blur as we discuss the topic, and with luck, we will come back to having to keep these 4 in mind
2) So what protections are available?
- Trade secrecy
contracts
copyright
patent
A trade secret is: a commercially valuable process.
It must have been initially secret; effort must have been made to keep it secret, and it must be commercially valuable
Liability comes from (1) breach of confidence; (2) discovery through “improper means” (not reverse engineering; but flyover in an airplane might be)
Starting in the 1950s; s/w producers relied upon trade secret; courts held that the source code is secret, but releasing the object code, qualified as a trade secret
Weaknesse:
decompilers are getting better and better (reverse engineering)
no protection against the piracy of object code
limited production against downstream consumers
consumers WANT the source code (forfeit trade secret protection if you give it out)
Contracts:
don’t sell, license the software - extract promises not to do things like:
- cannot resell
- cannot rent
- no claim of warranty
- no reverse engineering
- overrider fair use
- no contest of patent of copyright claim
Are these legal? Certainly challengable and lots of tussling. Largely upheld, although there are some limits
Pro-CD v Zeidenberg; Bowers V. Baystate Technologies
UCITA tried to embed this; largely as a result of law professors, this was stopped; adopted however in some state jurisdictions (Virginia)
Copyright Law
1964: software can be registered; but object code versions might not actually be copies
1978: CONTU recommend full copyright protection to software
1980: Congress adopts
1992: EC Directive 91/250: adopts s/w copyright
1994: TRIPS Agreement Article 10(1): adopts s/w copyright
Entitlements
-reproduction
-derivative works
-first distribution
-public performance
-public display
Exceptions:
-fair use
-merger doctrine — says that an idea cannot be protected, only the expression thereof. If there is only one way to do something; then you can’tprotect
-essential or archival copying
Apple v. Franklin - CA3 1983
Apple builds computers and also builds operating systems; Borland was given free access to create applications
Franklin computer company also makes computers; but sales depend upon running apple programs
So they copy the operating system;
Apple sues - Franklin made a copy
Franklin - we just copied the Apple o/s which was the only way that we could embody the idea of an apple O/S (merger)
Court decides - sorry, you have to make your own; (nice try!)
Over to Jonathan:
Alright, so now you know about the protection methods; let’s dig a little more into copyright, by starting about GNU (What do the Lion King and a penguin have in common?)
A cultural war afoot about copyright; many battles about understanding how creative work should be treated;
Let’s talk, then about copyright and software;
1969 - AT&T creates UNIX; did AT&T own UNIX? possibly, but Bell Labs decided to share the develoment
1974 - Berkeley UNIX (BSD)
1982 - BSD as the basis for Sun Solaris
Leading to a terribly complex software ecosystem with many parents and many versions
1984 - important year; AT&T signs a consent degree in the AT&T breakup of longdistance business. As a consequence, AT&T could now compete in things that it had been previously required NOT to compete within.
Richard Stallman decides that a new version of UNIX is needed; forms GNU - GNU’s Not Unix - see the gnu.org homepage
OTOH, GNU is supposed to be “a” unix, but not AT&T unix; how to create an o/s that works like Unix, but it isn’t unix — that’s the goal - to make that unix
Approach, keep replacing bits until Unix becomes all non-proprietary
2004 - state of play - features
Public domain, free (copyleft), Proprietary
Public domain
- works (edible cake)
- source code, for the most part (see the recipe)
- no license terms
- development model hard to perceive
Free
- - works (edible cake)
- - source code - the defining element (GPL) (recipe)
- - licensing terms - the general public license (more generally, the ‘copyleft’) (to block enclosure; conversion to proprietary)
the general public license — you can modify or copy the program and distribute it; provided that you put everything under the general public license (thus, the source has to come along with it, since that’s part of the license)
(Note that if you don’t redistribute, you don’t have to make the source public)
You can still charge for the physical copy of the work
-
- open development model — comparable to a barn raising
not entirely random; there are systems that have sprung up to organize to develop
Proprietary
-works (edible cake)
- NO (more or less) source code (no recipe)
- license - there are many, largely stating that all you get is the object code (the cake), but you can’t do anything else without asking
- closed/firm-based model
Is it credible that a bunch of randoms can get together to build something as complex as an O/S? Hard to say, but some things to seem to have worked.
Eric Raymond evangelized the idea of open source without scaring people (stop talking about “free’; stop talking about “abolish copyright”) talk about open source development as one that builds better software - successfully managed to make open source/GNU not a scary challenge to the prevailing ideology
So, we get SCO as a player in this game. SCO would sell Linux (rather than GNU) (Stallman wants GNU/Linux — long story) - very excited
Major disputes in software - back to Terry
Recall we ended with copyright.
3 varieties of s/w; public domain (no protection); free software (uses copyright to stay free); proprietary (depends heavily upon protection)
As copyright law evolved during the 1990s, it evolved into ways that were not encouraging to developers - talk about 2 in particular
Reverse engineering — different firms have adopted different strategies to deal with interoperability
1) free access - enough access to how the o/s works to let developers write applications
2) license fees - not generally available ; only provided to licensees ; lockout programs used to keep licensees from running applications (game hardware makers) — third parties elect to reverse engineer the lockout schemes (by copying the program that achieves lockout, and figure out how to circumvent it) and then their games will run (?copyright infringement?)
courts decide that this is essentially “fair use” — a strain to call it fair use, but the courts have done so - Atari 1992; Bowers 2003 (CAFC)
Ineffective enforcement - doesn’t seem to be terribly effective
Note, the statistics that will be shown are from the Business Software Alliance - generated via some ugly statistics, assume that all computers run the same software and any imbalance between hardware and software sales must be because of piracy
World piracy rate around 40%, more or less; regional US(25%) eastern europe (70%). If you look at $$, the losses are largely in Asia and the US - a list of shame (you can get it from the BSA WWW site)
Within the US, the piracy rates vary by region. Mid Atlantic is low, south central is high. BSA thinks it has to do with population density — the higher it is, the lower the piracy rate (odd?)
Terry has an alternative example — Blue v Red states (Gore v Bush) or above and below average piracy rates — interestingly, Red states (Bush) correlate with extent of votes for Bush *laughter*
So copyright seems to be evolving into something that do not give software developers the controls that they would really like
Back to Jonathan:
The SCO v IBM case (another form of internet gambling — possibly gambling with the future of the internet)
Recall, SCO and linux were working together. That was the old-SCO; the new-SCO is a little different. “We’re developing new and innovative licensing schemes.”
How did we get here? Recall the history of Unix; long path. In 1986, IBN makes AIX under license with AT&T. GNU Linux emerges out there supposedly independent from the AT&T codebase.
Novell decides to get Unix from AT&T. SCO emerges from the shambles of Novell’s unix foray with the rights to AT&T’s unix (1995)
Caldera emerges as a GNU/Linux consultancy. (1995)
The old Caldera and SCO bonded (into Caldera) to bought the AT&T rights; and then Caldera renames itself into SCO.
In 2000, IBM starts to contribute to GNU Linux.
And SCO says, wait a minute; the stuff you contributed, either from AIX or based on your original AT&T license, is owned by us - and now Linux is tainted because it has stuff that we own in it.
SCO is not just blaming IBM in this. They also argue that the interoperability efforts (done by copying source at the outset and then running to reverse engineer) also make Linux tainted. SCO says you stole stuff to make Linux (can’t say who, a bunch of randoms). We’re going to sue end user
Also embedded applications of Linux, like TiVo. We want $32 dollars for each one - why 32 - who can say?
Consultants offer up suggestions to “assess your exposure” - not terribly productive; in fact counterproductive.
SCO says they are going to sue an end user — pay for the amnesty ($600-700; plus accept our restrictive license) or you might be the one that we sue.
SCO ups the threatening letters; targeted to each end user (e.g. Lehman Brothers shown) Good lawyering, says please talk to RedHat, and stop bothering us.
Online attention - the SCO countdown
Targets — AutoZone (and Daimler-Chrysler)
What’s really going on? Note the rise in stock price with the announcement of the suit; peaking when Microsoft forked over $10mm (estimated) to get a Unix license. Microsoft has disavowed any “strategic” intent in offering up this money.
SCO - a company whose fundamental premise is based on lawsuits.
(Note that copyright has driven this discussion to date)
Back to Terry: and we’re now going to look at software patents
With the weakening of copyright and s/w; a new system emerges. Patent law.
Originally, patent law comes in two flavors: product and process. This keeps expanding - designs; pure substances; genes; special statutes for plants; animals; surgical procedures (revoked?); software; business methods
List of relevant cases; Gottschalk v. benson; Parker v. Flook (NO)
Diamond v Diehr (maybe if machine); turned over to the DC circuit, who keep relaxing the rules. State Street Bank — repudiates the business method patent as well
Lots of these patents - over 100,000; US leads, but other jurisdictions catching up.
EU now also seems to be on the verge of accepting this notion.
Intriguing subtleties; but we need to jump to the fundamental question?
“What sorts of IP Protection (if any?) are appropriate for software?”
How should the legal system mediate the battle between the open source and proprietary s/w world? What protection should be extended to proprietary software.
Most economist say the justification for extension of patent when these 6 are met
1) high R&D costs
2) high uncertainty of success
3) content of advantages can be figured out by competitors
4) advances can be easily mimiced
5) money is a strong, necessary stimulus to creation
6) innovation is not cumulative
For example, pharmaceuticals (possible except for 6) meets all of these.
Software is on the weak end of this measure
1) it’s cheap
2) you know you’re going to make something workable
3) reverse engineering is hard; decompilers not that good; lock up the cource
4) yes - this is true
5) really? what about GNU? Yochai’s talk
6) definitely cumulative
So, as far as economic theory says patents are a bad idea; empirical evidence tends to support the thought that patent protection is not a core element if software business/innovation in this area.
So, SCO’s struggle using trade secret and copyright to defeat open source may be depending upon a theory that doesn’t really obtain in the real world;
Charlie: Starting in 1984, the US IP crowd starts working on using the trade treaty strategy to extend US copyright/IP rules to foreign jurisdictions (TRIPS emerges and is now an international strategy, with Microsoft at the base of it)
Moreover, the notion of open forms has been around software; and many governments have publicly suggested they should support this; employ these tools instead of paying the Microsoft tax; a US company. Maybe governments should be putting more behind open source.
At each point, the Microsoft response has been to (1) tolerate piracy — network effects/lock-in; (2) lobbying directly and through State Dept against open source and (3) offering discounts and giveaways to keep Microsoft tools in use.
Education in particular seems like a place where the ideas of open source/free/collaborative development — this seems like a no-brainer; yet this is hard to promote and proprietary interests keep pushing this idea into the background.
Z: Yes, the absurd is the profound; it helps us to think harder.
We incentivize firms to promote their interests — that’s their job — we can’t demonize Microsoft for doing what comes naturally. What we need is government and public interest to channel their competitive urges into other avenues that are more productive for society
Luis from Ximian/Novell: To clarify, I’m a GNU user and developer. Yes, Novell keeps turning things into trash, but they do seem to be considering a change.
First demonizing: Microsoft is sponsoring the Irish president’s (www site?) — Irish Presidency sponsored by Microsoft (note that the Irish president is the holds the revolving EU presidency)
Second, the EU Council is going to push for s/w patents over the vote of the parliament
My experience is that the ideology of GNU is more accepted overseas; why don’t we use it more
Tim Granton from Microsoft: set aside the demonization - I think a lot of the debate is framed into the binary way - free or open/public or private — the polarized nature of this gets in the way of getting things done. The question is what’s the appropriate mix - and what soft of IP will give you the kind of innovation that you want. What IP protection scheme does a country want to set up to get development in these jurisdictions — that should be the focus
Terry: Followup questions? Want to talk about shared source? And why is Microsoft so aggressive about patenting software?
Tim: Sorry, don’t know enough to comment on shared source. And I’m not a patent lawyer, either. From a corporate perspective, when you examine your assets, you want to protect them. I would disagree with the economic analysis; there are some things about software that means that there should be patent protection.
Copyright is not working terribly well, though-, particularly with the Internet
Q: Is the alterntive compensation system applicable in this situation?
Terry: Probably not; software is far less fungible than music, so it’s really hard to use simple tracking technologies to value the IP.
George: Why patent protection is more attractive? Looking at the 6 point list, the first item (particularly the need for human factors/interface/etc) suggest that you have to pay a lot to get stuff to work
Q: If you are trying to protect something like software, where innovation is cumulative, unlike movies or music. Is there a way to differentiate betyween the two so that Valenti doesn’t end up being a proxy for Bill Gates?
A: Looking at the economic list; and assume that some of the first elements do obtain, what might you do? You could define specific legal instruments that would control basic release (protect against piracy), but would free up derivative works.
On the tactic, we are seeing the success of a deliberate campaign to turn copyright, patents etc from “limited monopolies” into “intellectual property” — and we conflate copyright with property rights. It looks like we’re past the point of no return on that issue.
Finally, the global campaign that Charlie raised. Yes, it has been vigorous and successful, particularly in developing countries — the so-called TRIPS+ — people are being asked to sign Title 17 entire to get foreign aid, trade, etc. — even though most provisions are completely irrelevant.
On the other hand, the enforcement of these rules is not terribly good; and it may be a deliberate government strategy - a classic pattern of helping local companies to develop products that can be sold in zones of high copyright protection (maximizing rent seeking)
Z: We have lots of discussion pent up, but we’re out of time……