Should Governments Specify Licence Conditions?

I have been made aware of a meme passing around Government purchasing circles to the effect that Government ought not to be dictating licence terms in the course of procurement.  This has two variants, a strong variant that Government ought not be specifying, for example, a class of licence that ought to apply to the procurement and a less strong variant to the effect that Government ought not be specifying particular licence terms. Of course, the underlying aim of this meme is that if a Government can’t dictate licence terms then it can’t require open source.

To argue these positions requires a complete lack of understanding of the role that a licence plays in an acquisition.  I will take software as an example, but any procurement involving a licence would serve as well.   When anyone “acquires” a piece of software they, primarily, acquire two things.  The first, is a copy of the software being acquired.  The second is a licence in relation to that software.  Neither is useful without the other.  A copy, even legitimately acquired, can’t be used* without a licence and a licence can’t be exercised without a copy.   However, of these two components – the licence and the copy, the licence is by far the more important because it demarcates the whole of the uses to which the copy can be put.  If your licence is good enough, you can dispense with the provision of a copy because you can acquire the copy from elsewhere.   The acquisition of the licence, and the terms of the licence are the greater part of the substance of the procurement.

To take a practical example, if I were to buy a copy of Office from Microsoft I can choose from Office Home and Student 2013 or Office Home and Business 2013.   Microsoft provides a comparison chart which discloses that the main difference between these two packages is that the first can only be used for “Home Use” while the second can be used for “Home or Business Use”.  Now, the purpose for which I might use Microsoft Office is not a function of the copy of the software I acquire.  It is wholly derived from the licence terms which apply to that copy.  To argue that the Government is not able to specify the characteristics of a licence is to literally prohibit Government from discriminating between a licence which permits only home use (which would be useless to the Government) and one which permits use in the course of business.

For a public servant to even entertain the possibility of a broad based limitation on specifying licence characteristics would be to demonstrate a total failure to understand the subject matter.  The licence is the substance of any software acquisition.  To not be able to specify licence characteristics is equivalent to not being able to include technical specifications in any other sort of acquisition.  It is a nonsense.

The only time where specifying a licence ought to be prohibited is where the licence terms effect an exclusionary dealing.  So, if the licence terms permitted use only by persons who had signed up for some form of online service being offered by a third party, that would be anticompetitive because it would require bidder’s  customers to be funneled through to the a third party.  Open source licences do not have these dependencies.

* technically, some uses may be permitted if they do not involve an infringement.  However, the scope of things which count as an infringement these days is so broad that in any practical scenario the use of software will involve performing an activity which would, in the absence of a licence, infringe copyright.

Corporate Structures for Free Software Projects

Corporate Structures for Free Software Projects

Q: I have some software which I would like to form the basis of a GPL project.   Should I set up an organisation and vest the copyright in that organisation.  Would it make enforcement easier?

A: This is basically a difficult question and a lot depends on the circumstances.  Some quick comments are:
* in theory, anyone who has contributed code, the copyright in which they hold, to a GPL project has standing to sue.  The damages they can recover may be small if their contribution is small.  However, the Act has provision for an award of additional damages where an infringer has been flagrant (etc) in their infringement.  So there is still scope for recovering damages where actual damages are small.

* if an individual takes action they are exposed to the possibility of having to pay the other side’s legal costs.

* vesting copyright in an organisation makes the copyright an asset of the organisation.  If the organisation is ever bankrupted (eg having to pay some other party’s legal costs), it would be one of the assets distributed among the organisation’s creditors.

* having assignments in favour of a single aggregating organisation will allow the organisation more flexibility in dealing with the material.

* having a separate entity will bring with it an additional administrative overhead and involve tax consequences.

* an organisation might be required by potential contributors or, equally, might not be acceptable to potential contributors.

* managing the assignment process and record keeping involve effort.  It may discourage contributions.

 

 

Open Source Law Releases Report On Open Standards

Open Source Law Releases Report On Open Standards

Update: UK Open source principles released

I have been doing a bit of work for a variety of people recently relating to standards and standards setting.  In early May I saw that the UK open standards consultation process had been extended because of a potential conflict of interest by one of the facilitators.  Linux Australia commissioned a report from me about Open Standards.  That report (link below) was completed last week and, I understand,  Linux Australia has used it as a basis for a submission to the UK Open Standards Consultation process.  The report covers a variety of issues relating to open standards.  Some of the issues it covers are:

• the difference between open formats and open standards.  I think if government focusses on open “standards”, then that’s a big problem.  In practice it would resolve to “open standards or anything goes”, with no middle ground.  However, there are plenty of viable open formats which have not been standardised.  Moreover, invariably any format used by an application will initially not be standardised, so there needs to be an incentive for that first format to be open.  The report suggests, for example, that open formats should be preferred to closed standards;
• what is open in relation to a format.  The word “open” is tossed around in government with varying degrees of precision.  Often it can mean no more than “specified”.  Thus, a format which requires the payment of extensive  licence fees can nevertheless be “open”.  The report endorses the proposition that if some format gives a preference to a particular bidder, then it’s not open.
• discriminatory “non discriminatory” terms – the report draws attention to the fact that licensing terms for standards are commonly overtly discriminatory.  For example, the requirement to pay per copy licence fees directly discriminates against any open source implementation of the standard.  Despite this obvious bias, such terms are routinely permitted to be categorised as “non-discriminatory”.  The report rejects the usefulness of terms like “RAND/FRAND/RANDZ”.

The report is licensed under a creative commons licence.
The full report can be downloaded here.

Here are the figures:

Figure 1: Govt Cost Shifting

Figure 2: File Format Feedback

Figure 3: Format Grid

CeBIT Gov 2.0 Conf: Open Source Wrongness

CeBIT Gov 2.0 Conf: Open Source Wrongness

I was contacted earlier today by “Stephen from CeBIT”.  Stephen ultimately was asking whether I would like to pay CeBIT for the privilege of presenting on open source related issues at an upcoming Gov 2.0 conference.  Stephen’s line was that there would be many potential customers at the conference so it would be a good investment.  In effect, Stephen was asking to be paid for marketing to Government.

Selling to Government, at least in Australia, is universally acknowledged to be difficult for SMEs.  Ultimately the reasons for this are that the Executive is particularly averse to failure and are subject to fairness tests in the award of contracts.  As a result, the Executive establishes a bureaucracy to ensure that each potential supplier is treated the same, and any engagement is subject to particularly extensive terms and conditions.  All of this carries with it a cost of engaging with Government.   In other words, marketing to Government is particularly expensive.

This is a particular problem for businesses based on open source because it means that the costs are heavily front loaded.  Part of the reason businesses pursue an open source strategy is that they do not have a marketing budget sufficient to kick start their operation – they may barely have enough in order to develop the code.  Several years ago John Roberts then CEO of SugarCRM spoke about marketing leverage of closed source vs open source businesses:

“I started looking at the financials of proprietary firms, and I started seeing that, some of the largest CRM providers that spent 80% of their operational expenditure on sales and marketing, and less than 10% on engineering.”

In other words, for an open source project to compete in a marketing sense, they would typically need to increase their headcount by a factor of 10.  Note that none of these extra people are improving the value of the software or solution to the Government customer.  All of them are engaged in lobbying the Government customer to adopt the product.

It gets worse – the expectation of Government is that these marketing costs are front loaded, and that they be recovered through licensing fees.  As mentioned above these costs are extraordinarily high in a relative sense for open source SMEs.  The corresponding risk of bidding for work is therefore very high – and higher the smaller the SME.  The fees for the provision of services are seen to be largely undifferentiated – an hour of service on product X is seen as being roughly the same value as an hour of service for product.  However product X is seen to be not comparable to product Y almost by definition (as their feature sets are different and/or the products may have network effects through lock in).  The upshot is that no premium can be charged for the price of services, but a premium can be charged for the grant of a licence. Thus the only way to recover these costs is through the licence fee.  You can see why this would be a problem for open source businesses.

It is somewhat pointless pursuing Government engagement with open source on these terms because genuine open source bids will always be underrepresented.   The Gov 2.0 conference looks to be an exemplar of this problem in microcosm.

Government instead, needs to be proactive in seeking out and evaluating open source solutions and, in particular, being technology neutral in its acquisition terms.  The Gov 2.0 projects seem to be not a bad model – with funding provided against open proposals with a comparatively low engagement cost.  Standardising on  closed data formats is particularly unhelpful, regardless of whether the format is an ISO standard.  If necessary, Government needs to reengineer its procurement practices as necessary to address any procedural fairness  issues.

2011: The Year of the Linux Desktop

2011 is the Year of the Linux Desktop

Hah! Not really.  I’ve been reading two posts, the first by Robert Strohmeyer, the second by Steven J. Vaughan-Nichols.  Both raise arguments about Linux on the Desktop and both point to mobile computing as being the future.

Ever since Android has come out I have assumed the growth path of Linux (and the ultimate strategy of Google) will be Android on phones -> Android on desktops.  My take on the Netbook episode is that, where customers returned Linux netbooks they returned them because they were unfamiliar.  With Android now in everyone’s pocket they won’t bat an eyelid at Android powered tablets (which I doubt were in Google’s game plan, but given that Android is open, others are  now able to fill that void), then Android netbooks and laptops and finally desktops.  With penetration of Android will come mobile developers and with them will come a large application suite.  Those applications will automatically run on an Android desktop.

On the mobile side of the world, I can’t see a mobile device replacing my desktop anytime soon.  However I wouldn’t be averse to a high level of integration between my mobile device and my desktop.   Indeed, as a user, and particularly as an IT Manager, I will probably see the benefit of having a consistent user interface across all my devices.  For this to happen either my mobile device could become Windows or my desktop could become Android.   I think the latter will be the easier transition, given that it is easier to move from an interface designed to cope with device limitations to a more capable device than to move in the other direction.    It is for this reason that I think it’s too early to write off Linux on the Desktop (LotD for Dohn Joe’s benefit ;-) [1].

The LotD Play is not one which anyone is used to.  There is no company betting it as a make or break decision, and even if there is (Canonical?), if they are broken, they are just part of the ecosystem, others will take their place.  That is to say, there is no lynchpin in the LotD ecosystem, without which it will fail.  This is what makes it different to the other operating system plays which have been out there.  If the guiding company couldn’t make its profit targets or satisfy its shareholders/investors/bank managers, it was curtains for the company, and by extension the technology.  Not so  LotD.  Like Obi Wan, should Vader strike it down, it will only become more powerful than he can possibly imagine (Linux on netbooks, for example, has become Android on phones, and need anyone forget the SCO debacle?).   If any LotD player falters others can take their place.  Moreover, they can take the benefit of the work already done and do not have to reinvent the wheel.

Finally, I think that another of the main difficulties faced by LotD is the lack of a level playing field.  The world over, legislatures (and history will judge them harshly for this) have been happy to pass laws which make people fearful of sharing.  Equally, governments have been particularly biased against open source offerings, although that bias is typically implicit in that they fail to implement open standards, or require open source to work within a procurement framework designed for closed source acquisitions.  Despite these obstacles the ecosystem which has the Linux kernel at its center continues to grow.  Governments are slowly removing bias from their procurement practices (some as a result of the pain of the GFC), and more and more agencies are independently implementing open source solutions.   LotD is the logical endpoint.

As I have argued elsewhere, I think there is a shift in the undercurrent which is pushing computing towards LotD.  I wouldn’t write it off now.  I wouldn’t write it off ever.

[Update (1 Nov): Overheard in a coffee shop this morning:

P1 (on phone, but to P2): What’s it called?

P2 (Beside P1): “HCC Desire”

P1 (to caller): “HCC Desire.  H… C… C…”

P2 (getting HTC Desire out of pocket): “Oh, H Tee C”

P1 (to caller): “Sorry, H Tee C – T for Tom.  It’s like an iPhone only better.  Can you get one? Ta.”

Notes:

1. Although after watching 10 years of such predictions I am wary of saying it will happen in the immediate future.

Open Source Licence Non-Compliance == Legal Trouble

Open Source Licence Non-Compliance == Legal Trouble

Brendan Scott, September MMX

For those of you who haven’t seen it, I have recently released the results of some research work conducted into the Trade Practices Implications of Infringing Copies of Open Source Software.   Linux Australia has agreed to contribute some funding towards this research note.  This is Australian law specific.

The main finding of the research is that a corporate vendor selling an infringing copy of open source software is likely to be in breach of at least one section of Part V the Trade Practices Act 1974 (Cth) relating to misleading or deceptive statements or conduct, and likely more than one. There are many cases in which such breaches have been found in relation to infringing copies of software. Even where a vendor only offers to sell (as opposed to actually selling) an infringing copy they are still likely to be in breach of the Act.

The Research Note is available here: Research Note on Trade Practices Implications of Infringing Copies of Open Source Software and also from the publications section at opensourcelaw.biz.

Bilski (on sware patents in the US) is out

Bilski (on sware patents in the US) is out

After weeks of “still no Bilski decision today again” stories in the press, they’ve handed down judgment and apparently software patents aren’t in as bad shape as they might have been.  Although from a look at secondary sources, they aren’t necessarily in good shape either with the Court pretty much avoiding the real question.

Decision

Wikipedia

SFLC

Luis Villa

Release of IFOSSLR Vol 2, Issue 1

Release of IFOSSLR Vol 2, Issue 1

The next edition of the International Free and Open Source Software Law Review is out [disclosure, I’m on the editorial committee] and it’s a great read.  I can wholeheartedly recommend the articles in this edition.  I wasn’t a reviewer of any of them for this edition so the first time I saw them was when I downloaded this edition.  They’re good.  It’s pretty rare I pick up a legal journal and add pretty much every article to my ‘to read’ list.

Moreover, even if you’re not a lawyer, this edition has some things which will be of practical interest to you.

Get it here.

The Patent War of All Against All

The Patent War of All Against All

“Hereby it is manifest that during the time men live without a common power to keep them all in awe, they are in that condition which is called war; and such a war as is of every man against every man. For war consisteth not in battle only, or the act of fighting, but in a tract of time, wherein the will to contend by battle is sufficiently known: and therefore the notion of time is to be considered in the nature of war, as it is in the nature of weather. For as the nature of foul weather lieth not in a shower or two of rain, but in an inclination thereto of many days together: so the nature of war consisteth not in actual fighting, but in the known disposition thereto during all the time there is no assurance to the contrary. All other time is PEACE.”

Hobbes, Leviathan, Chapter XIII, paragraph 8

Glyn Moody has written a post about a system called U-Prove.  Glyn notes that the software is being licensed under a BSD licence and notes that is a good thing, but then observes that there is a patent encumbrance on the code, and indicates this is a bad thing.    In a comment, Sam Ramji refers off to this article of mine, kindly remarking it is a ‘good post’.

The essence of Glyn’s argument is that the OSP does not preserve the freedom of free software, so it is not sufficient.  Sam appears to be adopting my comments to the effect that it’s hard to single out one company for the patent blame game.

In the context of free software patents are problematic.  In the ideal world patents on software wouldn’t exist and there wouldn’t be a problem.  However, they do exist.  Moreover, part of the reason they exist is because of a variation of mutually assured destruction – many businesses believe they need to acquire patents in order to defend against other patents.*  Jonathan Schwartz sets out some of the sad, tawdry circumstances in which this logic plays out here.

I think it is a non trivial problem to find wording which preserves just the defensive potential of patents (which, is actually their offensive potential limited to specific circumstances of exercise)  while preserving freedom when licensing software.  Some of the more detailed free software licences attempt this.   It is, I think, a more difficult problem to craft such wording to apply to standards – because standards purport to be agreed by some collection of people, while freedom requires that everyone be permitted to pursue their own goals.   Thus, any ‘promise’ or ‘covenant’** which is limited to an agreed specification must necessarily be inconsistent with freedom in a way qualitatively different to a patent clause in an open source licence.   Moreover, any wording which applies to a particular version of a specification will be inconsistent with the evolution of that specification.  In short, promises made in relation to specifications are likely to always be problematic (the best to hope for is a disclaimer  – per W3C).

Therefore, if the words themselves are likely inadequate, the issue of who is saying the words, and what one can reasonably read into them becomes much more important.  In the OSP post that Sam refers to I explicitly reference  another, earlier, article on the OSP which calls this issue out.  With this in mind, I note that Microsoft has recently chosen to specifically draw out the Linux/open source angle in its cross licensing deals with Amazon[3] and IO Data[4].  It didn’t need to, but presumably chose to.  It seems reasonable to conclude this is signaling.  These are grounds that would justify a reasonable person finding the OSP inadequate.

If Microsoft wants people like Glyn to trust them, they should perhaps incline against, rather than toward making the will to contend by battle […] sufficiently known.

Notes:

* This, by the way, is much the same argument used by Hobbes in favour of the need for a common power to counter the war of all against all (hence the quote above).  In this case however the common power (parliament) rather than saving us from a state of war of all against all has plunged some of us into a version of it.

** (I distrust these words – if it is a licence why not call it that, but perhaps they are US terms of art?)

[3] ‘…said Horacio Gutierrez, corporate vice president and deputy general counsel, in the news release. “Microsoft’s patent portfolio is the largest and strongest in the software industry, and this agreement demonstrates … our ability to reach pragmatic solutions to IP issues regardless of whether proprietary or open source software is involved.”  The press release is here, it doesn’t fall over itself as much as the IO Data one below.

[4] ‘...said David Kaefer, Microsoft general manager of intellectual property, in a statement announcing the latest deal. “Microsoft has a strong track record of collaboration with companies running Linux-based offerings, and this agreement is a reflection of our commitment to partner with industry leaders around the world.” ‘ In fact, on Microsoft’s press release page (at 12 March 2010) the call out comprises of the words:  “I-O Data customers will receive patent covenants for their use of devices running Linux and related open source software.”  The press release is at pains to refer to it.

Free Software is Principled

Free Software is Principled

I recall, several years back now, being in some sort of forum somewhere arguing over the implementation of anti-circumvention legislation in Australia.  I recall Rusty Russell talking about ghostscript’s [?] handling of pdf documents at the time and how it respected restrictions settings in the pdf documents.  That is, despite being able to ignore them, ghostscript’s authors decided to respect them.   In practice that would mean that most ghostscript users would also respect those settings.

By way of contrast, today, looking for information about pdf to text conversion tools I came across closed source software whose primary purpose is apparently to remove restrictions from pdf files.  In my experience free software is typically more principled than its closed source counterparts – perhaps stupidly so.