recent DISCUSSION

• Carl Cargill says:

  All -

  Several points that bear on the discussion.

  I’m fascinated by the discussions about the periphery of standards and standardization. If you go back to a basic question – which I’ve not seen touched upon directly – from whence does the power and legitimacy of standards derive? I would assume that it is a governmental grant to private parties to exercise a certain power to shape the market – which is what standards really are, after all. They are no more special than any other management discipline such as accounting, finance, or any other of the myriad functions management can exercise to promote product – which is the ultimate rationale for standards (“we don’t do them for intellectual stimulation, we do them for product”).

  So, if the government has the right to regulate advertising, and accounting, and production mechanisms, why don’t they have the right to regulate standards? I imagine they might make a mess of it, but it wold certainly be no worse than the 300+ competing and non-interoperable consortia and the host of squabbling SDOs.

  The question is – what serves business better (to be blunt.) Open source software – and the entire open source movement -came about because of the egalitarian nature of the software industry. The IEC (which tends to deal with big companies) doesn’t see the value in consortia – but that’s because there aren’t but a limited number of participants in their areas of competency. There are millions of software developers – and it is this factor that has skewed the entire discussion. It’s open source software, not open source hardware or open source sanitary rules or open source heavy equipment.

  So, let’s confine the discussion to the market phenomena that led to the issue – how do you deal in a structured standards environment (whatever that is) with an suddenly democratic and larger user base that doesn’t want to “play by the old rules”. Is is a phenomena that will be solved by time and maturity? or is it the forerunner of social movement and not a transient phenomena?
• A. Rebentisch says:

  As our world is not black and white but gray the matter is not about strict rules and exclusion. We also know that SSOs often fail, governments often fail, market mechanisms often fail. It is about preferences, incentives mechanism which lead to more lighter shades of gray. When you think about Open Source and patents the pragmatic question is how to reduce legal uncertainty and trouble. But @StephenWalli: There are certain fields with special requirements, you cannot accept software failure in an air plane, a government agency interface may raise non-discrimination, multilinguality and accessibility criteria, for certain governmental applications disclosure of source code (> open source) is crucial etc. The open source models developed first under the historical conditions of *nix user environment where non-disclosure of source code reduced the ease to migrate to other hardware, reduced the ability to fix bugs and optimize the software etc. Other models led in the market to fragmentation and friction costs. Usually you had professional administrators of the software with development skills. Of course the average desktop PC or smartphone may set different conditions, user needs and challenges than the highly fragmented Unix market. Standards and OSS models are both means to overcome the fragmentation and inefficencies of a market. Most SME developers hate patents on standards – not because they are unwilling to pay: because of the contraction costs and uncertainties involved. Quite common in digital fields: market failure via commercial fragmentation (and complication) vs. free flow of information without (sustainable) renumeration scheme for contributors.
• Stephen Walli says:

  @A. Rebentisch: “It would be unacceptable to require a citizen or business to license a patent from a private stakeholder to interact with their government authority, ….” I totally agree. I was hopefully clear in pointing out that there are alternatives. If the government thought the system they were procuring to provide a service was the “best”, the government would need to absorb the cost. From a slightly different perspective, the company holding such a patent would legally have the ability (in the U.S.) to sue infringing customers, but no company sues customers.
• Rebentisch says:

  “move away from perfect competition to provide innovation incentives.” – Exactly, that is the core idea. As far as Europe is concerned there is no polical endorsement of software patentability. In the field of standards patents are even less desirable, to quote Article 6a of the directive draft “Member States shall ensure that, wherever the use of a patented technique is needed for a significant purpose such as ensuring conversion of the conventions used in two different computer systems or networks so as to allow communication and exchange of data content between them, such use is not considered to be a patent infringement.”
• Tim Simcoe says:

  @Rebentish says “In terms of inducing an RF allocation we find here a market optimum, perfect competition. From a market rationale the best allocation is unlicensed standards.” This argument assumes away the idea behind the patent system: move away from perfect competition to provide innovation incentives.
  —–
  Once we introduce patents, the big problem facing the gov’t bureaucrat in JBC’s compelling scenario is not the existence of a trade-off between price and quality, but the possibility of hold-up. That is, once the gov’t (as customer) is locked in to a particular solution, the price will go up, not because of any intrinsic technical advantages of that solution, but because the govt’s bargaining position has become worse. RF-licensing, non-assertion covenants, ex ante disclosure of royalty caps and patent indemnification are all efforts to grapple with this problem. Traditional RAND commitments only insist that prices remain “reasonable” (which is what I meant below by RAND is “just a commitment to make licenses available,” though as JimH notes, it also places limits on patent re-assignments and injunctive relief).
  ——
  Let me end by re-phrasing JBC’s question in a way that makes it clear that there is nothing special about setting price=0… “Do the respondents here believe that it’s legitimate for governments, in some cases, to induce (by buying behavior) or mandate (by regulation) long-term pricing commitments from ICT vendors?”
  —-
  Thanks to all for an interesting thread,
  Tim
• A. Rebentisch says:

  Consider the scenario of an egovernment service. It would be unacceptable to require a citizen or business to license a patent from a private stakeholder to interact with their government authority, that would translate into a private levy for a public service. The argument with the tax payer costs does not hold because expenses and taxes are seperated. From a procurement perspective your negotiation position always mirrors the interests of a vendor, of course no procurement officer is interested in higher prices and no sales person in lower prices. In terms of inducing an RF allocation we find here a market optimum, perfect competition. From a market rationale the best allocation is unlicensed standards.
• Stephen Walli says:

  @JCB: Nice question. I fear it will cause us to vote along the lines of our backgrounds, however, and not necessarily bring about a consensus picture as we each define where we believe government responsibility lies as an actor in the economy. For my part I have spent most of my career in the commercial world as both customer and vendor. I think when the government behaves as a large customer in a marketplace, they should behave exactly as a customer and not try to drive public policy through their own (possibly unique) buying needs. I happened to live in the POSIX space through the standard’s development and implementation while being both a customer and later a vendor of POSIX related technology. Admittedly there were almost no IPR claims to be made. The U.S. government put a FIPS in place (via NIST) as guidance for procurement, and put their money where their mouth was by putting a FIPS certification process in place. They were attempting through the 1990s to move away from GOTS to COTS products to save the tax payer money and so adopted a commercial standard. It was encouragement to procure wisely (i.e. save money) while leaving the door open to buying the best solution which might mean long term efficiency savings. There were ways to escape the FIPS “requirement”. It don’t think it’s legitimate for government to encourage a standard (regardless of IPR content) through procurement, and it looks somewhat strange to have a patent system in place to encourage innovation and its publication for the public good while mandating royalty free standards.
• James Bryce Clark says:

  One specific thread seems to join many of today’s posts. But it’s not clear to me that there is agreement.

  Do the respondents here believe that it’s legitimate for governments, in some cases, to induce (by buying behavior) or mandate (by regulation) the use of royalty-free ICT systems?

  Assume that for some government-related data service, there exist some really-good proprietary methods, and some less fancy but adequate freely-useable methods. In such cases, the government is faced with a trade-off choice, during the “build out” phase. (We’re seeing this in “SmartGrid” work right now.) How does our hapless government official, sitting in the corrida squarely between the bull and the matador, decide which is more important, for a given case? The desire to let innovators to collect their rate-of-return on inventions via royalties? Or the desire to make government-spurred data & data services available without cost to taxpayers? Regards JBC
• Stephen Walli says:

  @James Bryce Clark: Well said. +1
• James Bryce Clark says:

  @Stacy Baird: Thanks for your comments.

  You and some antitrust/competition economists see this in the same way: there’s virtuous competition AMONG standards bodies for the best licensure policies, resulting in a matrix of fitting appropriate policies to risks & needs. Which needs vary. The same thing applies to all those OSI licenses, which multiply like bunny rabbits. As do standards organizations.

  Every peace treaty project faces the risk of losing momentum, from an elaborate tiring preliminary negotiation about the number of chairs and shape of table.

  If we are lucky, there’s a clumping and refactoring around some best practices, and the most successful models. That filtering process leads to better … but not complete … voluntary market normalization of licensing terms & expectations.
• James Bryce Clark says:

  @A. Rebentisch: Agree strongly with some of your points.
  – If regulatory incentives are set right competition authorities do not have to intervene. There are smarter ways to overcome friction.
  – Model patent license terms would certainly be helpful. The IDABC contributed their EUPL, an OSS license reviewed by participants of 27 European jurisdictions.

  Governments are not only regulators, now: they are huge buyers and market-leaders. They should use their power for good, in ways that are market-aware, not market-killing … and they are just starting to wake up to this.
• A. Rebentisch says:

  H.264 licensing as a recent example. Anyway, markets should not rely on the competition authorities and their shotgun methods. If regulatory incentives are set right competition authorities do not have to intervene. There are smarter ways to overcome friction. For a start, while you have five relevant oss licenses, you rely on customized RAND licenses/indemnification/pledges. Internationally private law is not harmonized at all. The same applies to patent laws, the EU has not harmonized national patent laws yet. Multinational model patent license terms would certainly be helpful. The IDABC contributed their EUPL, an OSS license reviewed by participants of 27 European jurisdictions. That’s a process for a government agency, I wish they’d develop RAND and RF model terms.
• JimH says:

  @Stephen Walli, to your point the W3C’s royalty free policy supports such a position.
• Stephen Walli says:

  To Stacy’s point: In general, order is important. If the FOSS project comes after the standard’s creation then the FOSS project creators have the lead offered by the standards participants on IPR policy as to the potential licensing landscape. A community-based FOSS project would do well to choose the FOSS project IPR policy in such a manner that it is consistent with the standard’s community wrt license, contribution agreements, and possible use of a foundation. A corporate-based FOSS project might naively choose an IPR policy inconsistent with the standard’s participants, but they run the risk of being ignored. If the FOSS project was the original model out of which the standard was developed (e.g. ODF and OpenOffice.org), then again one has a reasonable understanding of the existing licensing landscape regardless of what license was on the FOSS project. The fact that the FOSS project is under consideration as the model tells participants what they might be signing up for.
• JimH says:

  @Goerge, I happen to agree with this point that @Rebentisch raises. To your point, however, there is a difference between ex-ante price fixing that I am fairly confident the DOJ has said in the past is per se anti-trust violation, versus voluntary ex-ante disclosure and negotiation on a bilateral basis with royalty cap which is entirely consistent with FRAND.
• George Willingmyre says:

  I mentioned the Oracle/Google litigation over use of Java in Android and Suns original promises over Java. I do not have a position on the merits one way or ther other. However I point to some posts that contend there may be flexibilities within Sun’s original commitments consistent with the current Oracle litigation.
• Ajit Jaokar says:

  Thanks Stephen, George.
• George Willingmyre says:

  Hello Ken, you use the Microsoft APIs as an example of ” not adaptable” Are there postitive examples of what you are proposing?
• Stacy Baird says:

  I want to reiterate Jay’s call for a taxonomy, or analysis of the numerous open-source licenses. Indeed, most view the Open Source Software world as defined by the GPL, BSD and Apache licenses, but the proverbial devil resides in the details – license restrictions are the issue. Many of the 70+ open source licenses are based on one of the three: BSD, Apache or GPL. Those who use these licenses have to understand what they are signing up for and how those licenses interplay with the use of standards. There may be merit to shouting from the rooftops “GPL bad – BSD/Apache good” (in my political career, the sound bite could define the debate). But understanding how one is ‘bad’ is critical when people are asking for governments to shape public poicy. To echo Anne and Daniel’s point, the ‘bad’ is in the restrictive nature of many open source licenses, not the standards policy – this is critical to understanding the solution to the perceived problem. I am to some extent amused by the charge that RAND is incompatible with open source (GPL, to be precise). Many things are incompatible with the GPL, even other open source licenses – What I want to shout from the rooftops is that these “problems” arise because of the nature of the GPL. For example, Wikipedia notes incompatibilities between the Apache license and GPL: “The Apache Software Foundation and the Free Software Foundation (FSF) agree that the Apache License 2.0 is a free software license, compatible with version 3 of the GNU General Public License (GPL). However, the Free Software Foundation considers all versions of the Apache License (as of 2007) to be incompatible with the previous GPL versions 1 and 2. It should be noted, however, that there is a one-way incompatibility between the Apache version 2 and GPL version 3 licenses, in that you cannot include GPL version 3 code in an Apache project without activating the requirement that the entire project be relicensed under the GPL version 3. (Footnotes omitted).” I read this and have to say I am happy I’m not a software developer trying to figure out which code I can use in my work. It’s not just standards. The GPL is a complicated license with many restrictions. Therein lies the problem. In the more familiar world of license agreements, people negotiate the terms they are comfortable with (on both sides). GPL doesn’t enable that give-and-take. It sets out bright line rules. Some are finding ways around those rules (e.g., as I keep noting, new metrics to determine royalties that would be compatible with the GPL), but the fact that the “problem,” if there is one, is the text of the GPL, not whether a standard is FRAND or FREE.
• George Willingmyre says:

  Thanks for your comment Rebentisch. That image of me holding the many differing electric plugs and sockets around the world conveys incompatibily at a gut level.

  On what follows in your post shows that we disagree on a policy topic.

  clip Regulatory assistance may be required to raise the bar, so that standards do not fail to meet market needs and SSOs standards lead to more market efficiency end clip from Rebentisch

  There are antitrust and competion laws that SDOs and participants in SDOs need to be aware. Yesterday I saw a DOJ action agains a company that participated in meetings where prices were fixed http://www.justice.gov/atr/cases/huang.htm
• Ken Krechmer says:

  I have previously noted that to be open (IMHO), a standard that defines a programmable interface must be adaptable (see http://www.csrstds.com/OpnStdsCallforAction.pdf ). Adaptability is a complex concept which includes identify, negotiate and select processes. These processes provide the ability to identify, negotiate and select specific modes/parameters of the standardized interface, i.e., with or without associated IPR. Where adaptable interfaces are available there is no conflict between open systems and open standards. Where adaptable interfaces are not available, the standard for a programmable interface is not open, and there is likely to be some restraint on the “openness” of the open source software.

  Consider Microsoft APIs (server or application) which have been legally identified as being in restraint of trade (i.e., not open). If these APIs were adaptable then the OS/application API or the server/server API could negotiate to determine what to select. When a Microsoft server identifies a Microsoft server then they might operate in a proprietary Microsoft mode. When a non-Microsoft server or application identifies a Microsoft server or OS, the non-Microsoft server or application could select to operate in a public mode. Notice that an adaptable interface still allows a proprietary advantage, and also allows a public option. Had this been done perhaps Microsoft would not have been fined $1.3 billion by the EU.
• A. Rebentisch says:

  The draft document is no final version but a draft from consultants, the IDABC programme is followed by ISA. Actually government delegates rejected the IDABC EIF2 submission because it was not strong enough.
• JimH says:

  A couple of points, and I apologize for the late submission to the forum…

  (1) OSS (open source software) is not the same thing as an Open Standard. The former is a licensing model while the later is an SDO that is at least open for anyone to join;

  (2) @Tim Scimone, “I don’t see how RAND is just a commitment to make licenses available.” A RAND (or FRAND) commitment is believed by many SDOs (and participating companies) as a license for at least the following reasons: (a) offer ad incertas personas, (b) waiver of the right to an injunction in rem, and (c) pactum de non petendo. There are other reasons, but those 3 are a good start;

  (3) @Ajit, there are several SDOs that provide licensing commitments to anyone regardless of participation in the standard, and those committments are provided world wide; and

  (4) there is a good example in GSM for the need for government regulation when comparing the US with the EU models… the US was handicapped by regulation that favored too much competition and allowed multiple standards, thus allowing the dominant producers of GSM equipment and infrstraucture (Nokia and Ericsson) to run away with the worldwide mobile phone market.

  Many Thanks, Jim.
• Niklas Lindblom says:

  As mentioned, transparency and credibility is key in licensing. So what is the best way of achieving this? Is there a risk of creating confusion with “too many” models, or does a wide variety simply facilitate developers finding a model that suits them?
• George Willingmyre says:

  I suuport the comment from Ajit about the potential value of real examples of success.
• Stephen Walli says:

  To Ajit: I agree that simplicity and transparency as you noted are key desirability in the discussion. FOSS projects (regardless of whether implementing a standard) should be clear about their expectations with respect to participants with an “IPR policy” designed through their license choice and contribution agreements. They can likewise ensure they firewall projects through involvement with non-profit foundations (like my employer). Companies wishing to promulgate work should be clear in what they’re willing to license out and how (e.g. Google’s recent patent license around WebM). This helps remove the ambiguity.
• George Willingmyre says:

  on the question of governments getting involved in standards policy, I see as a global issue in procurement and egovernment policies. See for example in the EIF http://ec.europa.eu/idabc/servlets/Doc?id=31597 Also the egovernment of India National Open Standards Policy proposal and the China original SAC proposal for a standards policy
• Ajit Jaokar says:

  Stephen walli says ..
  —
  A lot of the discussion in trying to tease apart licenses treads the space of “GPL Bad, Apache/BSD Good”.
  –
  This is true. I have also been guilty of the same! But I think as I said in comment to Steve .. people like simplicity .. they will choose the option where there is max precedence so that they dont have to reinvent the wheel .. thats the main issue I think! rgds Ajit
• A. Rebentisch says:

  I am sorry for the lack of respect, Mr. Willingmyre but my laughter was real… Let’s not mistake software licensing, patent licensing and minimum patent licensing requirements of SSOs. It is not worth to discuss that standards should be unlicensed in the rationale of a market system but SSOs lack the authority to enforce that rule, so RAND requirements are the worst best. As of compatibility of SSO standards licensing requirements with software licensing it indicates that the worst best is often not good enough. Regulatory assistance may be required to raise the bar, so that standards do not fail to meet market needs and SSOs standards lead to more market efficiency. Think of egovernment solutions, it is simply unacceptable to pay a private levy for interoperability with a public service.
• Tim Simcoe says:

  Steve’s post raises the interesting issue of standards-related IPR that is not essential, but nevertheless highly desirable. Economist would call such patents “proprietary complements.” Is one advantage of OSS licensing that it provides more clarity about the treatment of those kinds of patents, which do not fall under the purview of a RAND commitment?
• Ajit Jaokar says:

  Steve says
  —
  This is a complex legal discussion and even experienced intellectual property and licensing lawyers can easily lose the thread. But there are important policy questions that we need to consider as we work through this set of issues
  —
  Two thoughts come to mind(maybe for a future post)
  1) There may be a need to simplify the issues from an IPR standpoint for developers of open source sw. Ex if someone wants to take this route, how easy is it? or would the legal fees be too expensive and they may be forced to give it ‘for free’
  2) Are there any examples of ‘success stories’ or OSS and IPR? That would be interesting! kind rgds Ajit
• Tim Simcoe says:

  I guess my my question is whether *any* of these licensing models necessarily “conflict” with RAND. If so, I don’t see how.

  RAND is just a commitment to make licenses available. Many OSS models go further, and ask for a commitment to low prices (yes, there are many complex details, but isn’t this the basic point?). Both approaches are potentially useful ways to reduce uncertainty about hold-up and royalty-stacking. And both leave firms with a choice about whether to contribute technology, implement standards or take licenses.
• Stephen Walli says:

  Picking up on George’s idea: A lot of the discussion in trying to tease apart licenses treads the space of “GPL Bad, Apache/BSD Good”. The GPL does prevent one from distributing software for which one doesn’t have all the rights (section 7), but this assumes black-and-white discourse and the bright-line between them.

  As Mr. Mutkowski observes this would not prevent a company selling a product in the marketplace from paying royalties to an IPR holder where the product contains software covered by the GPL that may infringe the claims of the holder. We have seen this with the Microsoft dance. Microsoft claims Linux infringes its IPR. Company A (shipping a product that contains Linux software) and Microsoft sign a cross licensing deal of some form. Microsoft tries to message this as a proof point for the infringement, while Company A states that it is happy with the cross license but doesn’t believe Linux infringes Microsoft patents.

  We have seen a maturing in the free and open source community over the past couple of decades with respect to how to deal with patents in their midst (regardless of opinions on whether software patents are “good” or “bad”). Licenses have evolved (GPLv3, ASL 2.0, IPL->CPL->EPL), foundations have arrived, development communities pay more attention, and product companies tread as carefully as they always have based on their size/revenues. Indeed software developers have needed to mature their IPR views in general. The examples above merely show how the FOSS community has handled it.

  There are lots of creative ways to deal with the ambiguity of patents in the FOSS space, regardless of license, and regardless of whether developers are implementing standards or not.

  There are
• Stacy Baird says:

  Ajit:

  There are over 65 Open Source Initiative listed open-source licenses, the list is at http://www.opensource.org/licenses/alphabetical
• George Willingmyre says:

  I see many contributions using the term “Open standards” and “proprietary standards” I included a sentence observing the confusing multiple definition of an “Open standard” I reserve the term “Open standard” for one that meets process criteria such as ANSI’s essential requirement. I may be fighting a losing battle but this is also the formal position of ANSI http://publicaa.ansi.org/sites/apdl/Documents/Standards%20Activities/Critical%20Issues%20Papers/Open-Stds.pdf
• Stacy Baird says:

  James Bryce Clark, excellent overview, thank you. You brightly illuminate the market at work. Indeed, there are uncertainties, particularly noted in your final paragraphs, but as you have illustrated, OASIS exemplifies players in the market identifying solutions to perceived problems, setting ground rules that comport with the needs of certain industry constituencies – again – illustrating the market at work. Furthermore, you implicitly explain the evolution in industry by your reference to globally networked open systems as compared to closed telecom networks. In this multi-modal world, it is harder for a single patent holder to ‘play hard ball,’ or even control as ‘the only game in town’ in the standards setting arena. There is far more competition and diversity, even in the standards marketplace. And some (many) of those standards, as you note, are unencumbered by patent licenses (or in some cases I would add, licensees devise new ways of making the license work for the licensor, for example, new metrics to determine royalties). You describe well that the pool is big enough for a variety of swimmers, particularly because the swimmers have set reasonable (and evolving) rules for themselves. They didn’t need the pool manager setting out float lines to define the lanes. To put a finer point on it, there are standards bodies and standards that meet diverse needs, those who call for draconian reforms should heed.
• Ajit Jaokar says:

  Helen says
  —
  For instance, “defensive patent pools” work similarly to SDOs in that they aggregate a set of standards for public use with the intent that users can implement the standards without facing litigation; however, in the case of “defensive patent pools” member companies pay an annual fee to access all the standards.
  —
  I know this works quite well in the telecoms domain(and you also give examples of wimax etc), However, I have one qs for you. You mention ‘member companies’ .. but what about companies who are not members? This is a common complaint from the Chinese vendors in telecoms for example i.e. the incumbents have sewn up all the technology and the cost of entry for new comers is high. especially if they are a threat to incumbents(cheaper). what is your thought on this? kind rgds Ajit
• George Willingmyre says:

  I found a site that lists 75 versions http://www.opensource.org/licenses/category
• Ajit Jaokar says:

  Jay says:
  —
  There are currently over 50 open source licenses in use.
  –
  I did not know there were that many. A taxonomy will be good but I think most people take the 80/20 rule ie to me GPLv3 and Apache are representative of the two sides of the arguement but I agree that a comprehensive list would be good rgds Ajit
• Niklas Lindblom says:

  It is certainly true that the details are incredibly significant. One argument from royalty-free proponents is that confusion regarding licensing should be avoided; but is there any confusion created by the emergence of several differing licensing models? //Niklas
• George Willingmyre says:

  A theme in many of the contributions is that there are many “types” of licenses for OSS. Some of the licenses are consistent with “RAND-like” patent policies of SDOs. Some are not. It is all in the details.
• Mattias Ganslandt says:

  Welcome! All comments during this forum should be made on the Intro page “Questions…”. Please joint the discussion. /Mattias Ganslandt, Editor