Archive for the 'it law' Category

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.

iiNet: The History of “Authorization” and “sanction, approve, countenance”

iiNet: The History of “Authorization” and “sanction, approve, countenance”
– The missing mental element of the primary infringer

Update 2012:: my notes on Authorization have become a paper: Scott, B, Authorisation under copyright law and “the nature of any relationship” (2011) 22 AIPJ 172. (Although it should read AuthoriZation, but changed because of Thompson Reuter’s house style :-/

Exposure draft 1 September 2011.

In the 1987 decision of Hanimex, Gummow J commented at 285 that ‘[t]he evolution of the meaning of “authorisation” in the 1911 Act and the 1968 Act has pursued perhaps an even more tortuous course than the doctrine of contributory infringement in the United States’. The Court concurs with such statement, and it has become even more apt in the years following that decision. Despite the legislature’s attempt to simplify the relevant considerations pursuant to the Copyright Amendment (Digital Agenda) Act and s 101(1A), the law of authorisation has continued to grow more complicated and unwieldy, with a litany of competing and contrasting considerations, and with one statement of principle frequently matched with a contradictory one. The authority on authorisation has become a mire. 

Cowdroy J  at first instance para 358


The City of Adelaide case is not authority for including “permit” within the ambit of the meaning of “authorize”.

The history of the words “sanction, approve, countenance” show that there is more to an authorization than merely the acts of the person alleged to be authorizing.   In addition to those acts, the mental state of the person carrying out the primary infringement is a necessary ingredient.  In particular, the primary infringer must have some belief that the authorizer has the right or power to grant such an authorization.  Acknowledging this component of “authorization” will make the various decisions much more sensible, and less miresome.

[While not argued here, we suspect that for this reason the decision of the judge at first instance should be followed in the iiNet case.  That is, in the minds of the users there is no possibility for them to have a mental state that iiNet is authorising them.  Cowdroy J’s focus on provision of the means of infringement is identifying the basis on which such a mental state can be supported.   If this mental state in the primary infringer is conceded, Cowdroy’s approach is not inconsistent with s 101(1A). ]

Note: this draft is made available for comment.


The High Court has recently granted special leave to appeal in Roadshow and others v iiNet (the iiNet case). This case is concerned with the liability of Internet service providers for authorizing infringements which occur through the use of the Internet service provider’s infrastructure. Under the law relating to authorization a second person can be liable for an infringement carried out by a first person in if that second person has “authorized” the first person’s infringement. The concept of authorization has been treated on many occasions in Australian case law. Its history is summarised by Gummow J in the Hanimex case. In particular, the High Court in the Moorhouse case held that a university library, by locating a photocopier near to racks of shelving containing books, some of which were still subject to copyright, had authorized infringements in relation to those works which occurred when people photocopied then using the photocopiers which had been so located.

In order for authorization to be made out, a “primary” infringement must be proven.  That is, simply saying to someone “go and make a copy of this” is not an infringement if they do not then go and copy.  However, if they do go and copy, then there are two infringements – the primary infringement, being the making of a copy, and the authorization of that primary infringement.  In the iiNet case this primary infringement is that of “making available online”. When a person “makes available online” a copy of a copyright work that “making available” itself constitutes an infringement.  In this case, the infringement is of the right to communicate to the public (defined in section 10 of the Act). Is in the lower courts this infringement has been established by users using peer-to-peer software and having the relevant copyright work within a particular folder for sharing by their peer to peer software.

Authorization isn’t – where is “sanction, approve, countenance”?

Unfortunately, in the Act, the word “authorize” has been interpreted to mean something other than “authorize”.   Rather, courts have read into the term “authorize” the concepts of “sanction, approve, countenance”, based on some words of Gibbs and Jacobs JJ in Moorhouse.

Moorhouse:  In that case Gibbs J says:

” The word “authorize”, in legislation of similar intendment to s. 36 of the Act, has been held judicially to have its dictionary meaning of “sanction, approve, countenance”: Falcon v. Famous Players Film Co. (1926) 2 KB 474, at p 491 ; Adelaide Corporation v. Australasian Performing Right Association Ltd. [1928] HCA 10; (1928) 40 CLR 481, at pp 489, 497 .” (Per Gibbs J at paragraph 10).

In addition Jacobs J says:

It is established that the word is not limited to the authorizing of an agent by a principal. Where there is such an authority the act of the agent is the act of the principal and thus the principal himself may be said to do the act comprised in the copyright. But authorization is wider than authority. It has, in relation to a similar use in previous copyright legislation, been given the meaning, taken from the Oxford Dictionary, of “sanction, approve, countenance”. See Falcon v. Famous Players Film Co. (1926) 2 KB 474 which was approved in Adelaide Corporation v. Australasian Performing Right Association Ltd. [1928] HCA 10; (1928) 40 CLR 481 . I have no doubt that the word is used in the same sense in s. 36 (1). It is a wide meaning which in cases of permission or invitation is apt to apply both where an express permission or invitation is extended to do the act comprised in the copyright and where such a permission or invitation may be implied. Where a general permission or invitation may be implied it is clearly unnecessary that the authorizing party have knowledge that a particular act comprised in the copyright will be done. (per Jacobs J at paragraph 9)

The third judge, McTiernan A.C.J. did not write a separate judgment, but agreed with the others. So this wording is essentially unanimous.  I note here that no dictionary I have gives “sanction, approve, countenance” as the meaning of authorize.

Adelaide Corporation: However, neither  Falcon nor Adelaide Corporation say that “authorize” has its ordinary dictionary meaning of “sanction, approve, countenance”.   In  Adelaide Corporation v  Australian Performing Right Association Limited, the  Falcon case was referred to in the dissenting judgment of Isaacs J who adopted something very similar to, but different from, the Gibbs/Jacobs  forumla:

“For the word “cause,” which had been given a restricted meaning, there was substituted the word “authorize,” and that word is now understood as importing the sense of “sanction, approve and countenance” (see Falcon v. Famous Players Film Co.).” [my emphasis – as we will see below Isaacs J is accurately quoting Falcon, whereas Gibbs/Jacobs JJ are not.]

Higgins J (part of the majority, but in a separate decision) also refers to the formula but in a contra-positive way .  Higgins J does not cite the Falcon case and does not explicitly adopt the wording as a test.  What Higgins J says is:

“This action has, by consent, been tried on the motion affidavits; and there is not the slightest evidence of any “sanction, approval or countenance” given by the Corporation to the performance of the song in question.” [my emphasis]

The other two judges in the majority in Adelaide Corporation (five judges heard the case, so even a minority so much as referred to the sanction, approve and countenance wording) base their discussion on the Performing Right Society v. Ciryl Theatrical Syndicate (1924) 1 K.B. ignoring Falcon and ignoring the “sanction, approve, countenance” wording.  Indeed, this is not altogether surprising because the focus of the judges in the City of Adelaide case was whether the Corporation had “permitted” the relevant activity.  Gibbs J in Moorhouse says Adelaide Corporation treated “authorize” and “permit” as synonymous:

“It can also mean “permit”, and in Adelaide Corporation v. Australasian Performing Right Association Ltd. [1928] HCA 10; (1928) 40 CLR 481 “authorize” and “permit” appear to have been treated as synonymous. “

However, this appears to be simply wrong.  This is discussed in the end note below.

It is hard to see that Adelaide Corporation is authority for Gibbs/Jacobs JJ’s statement.

Falcon: So, what of  Falcon?  That case references the”dictionary” wording as follows:

It has been said, and I think the comment is not an unfair one, that the use of that word “authorise” adds nothing to the definition, because, strictly speaking, the sole right of performance must necessarily include the right to authorize performance. But when the definition of “copyright” is read in connection with that of “infringement,” I think it becomes fairly apparent that the object of introducing the word “authorise” was to get rid of the effect of certain decisions, of which  Karno v. Pathé Frères  was the most recent, in which it was held upon the language of earlier statutes that a defendant who for reward gave permission to a third person to represent a play in breach of the owner’s copyright did not “cause it to be represented,” unless the person so representing it was the servant or agent of the defendant. In the present statute that language has been deliberately dropped, and for the word “cause” has been substituted the word “authorise”; and the decision of Tomlin J. in  Evans v. Hulton and the dictum of Buckley L.J. in  Monckton v. Pathé Frères  both clearly indicate that in the opinion of those learned judges the present expression is to be understood in its ordinary dictionary sense of “sanction, approve, and countenance.” If that is the true view I think McCardie J. was, upon the materials before him, amply justified in coming to the conclusion that the first three defendants had infringed the plaintiff’s sole right to authorize the performance of the play. For these reasons I am of opinion that the appeal should be dismissed. (per Bankes LJ at 490)

That is, in comments by a single judge  of the three judges that heard the case).  Note also here that the words used are “sanction, approve and countenance” (my emphasis).  These were faithfully reproduced in Adelaide Corporation, but were transformed into “sanction, approve, countenance” in Moorhouse.    The disjunctive form has survived in the  iiNet case – “The word authorise, when used in s 101, has its dictionary meaning of sanction, approve or countenance.” (per Emmett J at paragraph 24, emphasis in original).

Moreover, the Falcon case also had this to say about “authorize”:

For the purposes of this case it appears to me that to “authorize” means to grant or purport to grant to a third person the right to do the act complained of, whether the intention is that the grantee shall do the act on his own account, or only on account of the grantor; and that construction of the word “authorize” seems to have been the one adopted by Buckley L.J. in Monckton v. Pathé Frères , where he held that the seller of a gramophone record authorizes the use of the record, and by Tomlin J., who in  Evans v. Hulton held that a person who sold to a publisher serial rights in stories written by the plaintiff authorized the printing and publication of them. (per Atkin LJ at 499)

Evans: The reference to Evans v Hulton [1924] All E.R. Rep 224 appears to be the ultimate origin of the “sanction, approve, countenance” wording.  In Evans v Houlton the plaintiff, Evans, ghost wrote an autobiography of a fellow, Zeitun, based on an agreement that Zeitun would find a buyer for the story and would share the proceeds of any sale.  Zeitun told the plaintiff that he couldn’t find any buyers, then found a buyer, Hulton.  Hulton started publishing the story and Evans sued Hulton and Zeitun, claiming, in part, that Zeitun had authorised the publication.  The court found that, by providing the manuscript to the publisher Zeitun had authorised its publication.  In that case Tomlin J didn’t say that authorisation should have its meaning in the dictionary.  What he said was:

“It has been ingeniously argued that in the subsection “to authorise any such acts” means to sanction their being done by the servant or agent of the person affecting to give the authority on his behalf, and that there is no infringement of a copyright when that person to whom authority is given is not the servant or agent of the person affecting to give it.  In my judgment, this is to put too narrow a meaning on the word, which is defined in the Oxford Dictionary as meaning, in connection with the authorisation of acts, “to give formal approval to, to sanction, approve, countenance.” In my opinion, where a man sells the right in relation to a manuscript to another with the view to its production, and it is in fact produced, both the English language and common sense requires it to be held that that man “authorised” the printing and publication.” (per Tomlin J, Evans v Hulton [1924] All E.R. Rep 224 at 225)

I note here that Tomlin includes “give formal approval to”.  It is also important to note that Tomlin J gives the dictionary “meaning”.  in Falcon, Bankes LJ left out the first component, relating to the giving of formal approval to.  However, in Falcon Bank LJ doesn’t say this is the meaning.  Rather, he says that it has it’s dictionary sense of “sanction, approval and countenance”.   This might be considered a reasonable interpolation of the actual meaning given by Tomlin J.   Presumably, Bankes LJ felt the addition of “and” was sufficient to read out “give formal approval to” when proposing the formula he did.   Isaacs’ wording in Adelaide Corporation is consistent with this, referring to the dictionary “sense”.

The whole of the court in Moorhouse has purported to give “authorize” its dictionary meaning and then failed to do so.  The Moorhouse court adopted not even the dictionary “sense” averted to in the authorities cited, but  rather, have substituted a new formula entirely – that of “sanction, approval, countenance” (into which later judges have read “or”).  This almost completely upends the earlier authorities cited in support.

Monckton:  Falcon refers to the dictum of Buckley J in this case.  The judgment of Buckley J does not use any of the Moorhouse words at all, let alone in explaining the meaning of authorization.  Nor does it refer to ordinary or dictionary meaning.  In the case, the defendants legally manufactured a quantity of records and imported them into England.  Before they sold the records the Copyright Act 1911 came into force vesting in the composer of the works recorded on the records a right to perform the works recorded.  The court held that the sale of the records was an infringement of the right to authorize the performance of the works.  This seems to be the passage that Falcon is referring to:

“…it seems to me that the Act has given to the owner of the new and extended copyright as defined by the Act the sole right to authorize any one to produce the musical work (e.g., by the user of the record: s. 1, sub-s. 2), and has made it an infringement of his rights that a person should sell a record which to the knowledge of the seller infringes the sole right of the composer to produce the work by the use of the record: s. 2, sub-s. 2.. (per Buckley LJ, Monckton v. Pathé Frères [1914] 1 K.B. 395 at 405, Vaughan Williams L.J. “entirely agree[ing]”)

That is to say that what Bankes LJ is stressing in Falcon is the mental state of the primary infringer.  However, the Moorhouse decision has only taken part of the Falcon meaning.  Indeed, while the Moorhouse case does identify the need of a mental element, it is placed wholly on the authorizer.

Conclusion – The Primary Infringer’s State of Mind as the Missing Mental Element

What all of these cases establish is that the absence of service or agency is not intended to defeat a claim of authorization or rather, the purpose of the inclusion of authorization under the 1911 Act was to overcome a particular limitation found in the case law.  However, for all of them, the background of the discussion is one in which the primary infringement is a natural, if not necessary, consequence of the action of the authorizer and that this is known or understood by the primary infringer.   In Monckton, it was the expectation of both the seller and the purchaser that the record would be played after it was purchased.  In Hulton, it was the expectation of both the seller and the purchaser  that the manuscript would be published.  In Falcon,  it was the expectation of both the seller and the purchaser  that  the films would be shown in the theatre.  Further, in each case it was reasonable for the purchaser to believe that the seller had the power to allow the purchaser to do the relevant act.  [As an aside, it is likely that Moorhouse can be fit within this requirement, Cooper possibly, but iiNet unlikely]  Consider a thought experiment in which each of these cases was re-run, except that the seller was selling to a purchaser under a common understanding that they would be broken into its elements for recycling (to recover the paper in the manuscript, the shellac of the records and so forth).  Had the purchaser turned around after the sale and then published, played and so forth it is hard to imagine an authorization by the seller would be found in these cases, despite the seller having done the same acts.

In each of these cases there is not only a relationship between the authorizer and the primary infringer, there is a common understanding.   This is consistent with the explanation of the purpose of the wording being to catch third parties who are not servants or agents.  The summary of the plaintiff/respondent’s argument (not part of the judgment) preceding Falcon casts this into relief, as it explains:

“…[in  Evans v. Hulton], a person, who sold to a firm of publishers the serial rights in certain stories written by the plaintiff, was held to have infringed the copyright by authorizing the printing and publication. Tomlin J. there adopted the definition of the word “authorise” given in the Oxford Dictionary, “To give formal approval to, to sanction, approve, countenance.” That word is not, like the word “cause,” limited to the giving of directions to a servant. A man who has no legal right to do an act himself may authorize another to do it, if that other believes that he is in a position to give him the legal right to do it, which was the case here.” (at 486) [my emphasis]

The highlighted wording is significantWhile this wording does not form part of the judgment, it demonstrates the assumption of the copyright owner in pursuing the case.  Therefore, if the court is to give to “authorize” it’s ordinary, contemporaneous meaning, these submissions, coming as they are from the copyright holder, are relevant as evidence of that meaning.

This wording also identifies the common, but unstated thread running through all of these cases on which the meaning of “authorization” is based.    That is, to say, whether  a person A “authorizes” another person B to do an act is not purely a matter of the conduct of A.  In addition to the conduct of A, B must have a particular state of mind.  B must have some perception not only of authority flowing from A, but also of A’s right or ability to grant such an authority.   This is consistent with the “ordinary” meaning of authorization.

[ It is this mental element which is the reason Cowdroy J’s approach is likely the right one.]

End Note: Did Adelaide Corporation treat “authorize” and “permit” as synonymous?

The Adelaide Corporation case was concerned with section 2(3) of the Imperial Copyright Act (which had been adopted as law in Australia by our 1912 Act).   The subsection (3) reads:

“Copyright in a work shall also be deemed to be infringed by any person who for his private profit permits a theatre or other place of entertainment to be used for the performance in public of the work without the consent of the owner of the copyright, unless he was not aware, and had no reasonable ground for suspecting, that the performance would be an infringement of copyright.”

The Act uses the word “permits”, not authorizes.  The question put to the court was whether section 2(3) had been breached, so the court was not asked to form an opinion on the meaning of “authorizes”.   Now, reviewing each of the judges’ references to “authorize” in turn:

Knox CJ does not refer to authorize.

Gavan Duffy and Starke JJ observe:

Inactivity or “indifference, exhibited by acts of commission or omission, may reach a degree from which an authorization or permission may be inferred. It is a question of fact in each case what is the true inference to be drawn from the conduct of the person who is said to have authorized the performance or permitted the use of a place of entertainment for the performance complained of” (Performing Right Society v. Ciryl Theatrical Syndicate).

However, this is not to treat them as synonymous.  Rather, they are quoting a decision of Bankes LJ in Ciryl.  In that case:

The two questions to be considered were whether there was any evidence that the appellant (1.) authorized the performance of these musical works by the band of the theatre on the particular days or (2.) permitted the use of the theatre for his own private profit for the performance of the works on those occasions. (per Bankes LJ Performing Right Society v. Ciryl Theatrical Syndicate at 8)

So, it was appropriate for Bankes LJ to phrase things in this way because both issues were put before the court in that case.  Gavan Duffy and Starke JJ are simply quoting the wording of Bankes LJ (the quoted section appears in the judgment of Bankes LJ).

The only other reference to “authorizes” in the judgment of Gavan Duffy and Starke JJ is about the effect of a clause:

“Now, the clause does not give the Corporation any control over J. C. Williamson Ltd. or Hislop or over concerts given by them in the Town Hall: all it authorizes is a termination of the contractual relationship constituted by the letting agreement.”

Here authorization is not being used in its Copyright Act sense.

Issacs J refers to authorize (see quote above), but only for the purpose of illustrating the meaning of permits (and thereby distinguishing them):

“For the word “cause,” which had been given a restricted meaning, there was substituted the word “authorize,” and that word is now understood as importing the sense of “sanction, approve and countenance” (see Falcon v. Famous Players Film Co.[3]). Along with that significant change, there was inserted the sub-section now under consideration, which is a necessary complement in the scheme of protection. It extends the previous limited responsibility of proprietors of halls who merely hire them out for remuneration without being directly connected with the performance itself. The gist of the sub-section is that such a person who “permits” the performance of a copyright piece is to all intents and purposes as much an infringer of the copyright by providing for gain to himself the necessary means of infringement as if he directly produced the performance, unless he proves that “he was not aware, and had no reasonable ground for suspecting” that the performance would be an infringement.

Like the new word “authorize,” the word “permits” is of very extensive connotation. Having international effect, it should not be restricted to narrow limits. Going to the dictionary, as did Buckley L.J. (as he then was), for the former word, we find that in the Oxford Dictionary the primary meaning of “permit” is: “to allow, suffer, give leave; not to prevent.”

Neither of these references treats “authorize” and “permit” as synonymous.  Indeed, it assumes them to have distinct meanings.

Higgins J refers to “authorize” as follows:

“As Bankes L.J. said in the Ciryl Case, the indifference was “not of a kind to warrant the inference of authorization or permission. It was the indifference of one who did not consider it his business to interfere, who had no desire to see another person’s copyright infringed, but whose view was that copyright and infringement were matters for” others (here J. C. Williamson Ltd.) “to consider.” The essence of the position is that J. C. Williamson Ltd., had, on the existing facts, control of the performers as to what they should sing, whereas the Corporation had none.”

Again, this is merely a quote of Bankes LJ in Ciryl.  Higgins J also says:

“As the Supreme Court have also pointed out, to prove permission it is not enough to show that the Corporation gave a mere general authority for the performance of musical works, a “vocal concert”: it must be proved that permission was given to perform the particular works in question (and see per Atkin L.J. in Performing Right Society v. Ciryl Theatrical Syndicate). Where is there even a semblance of proof of such permission?”

But again, this is not a reference to authorization in the Copyright Act sense.  Indeed, it distinguishes the “giving of authority” and permission.  A review of Cyril, the case cited, also reveals a repeated reference by their Lordships to both “authorize” and “permit” as separate concepts, with no attempt to identify them as a single concept.  One can concede that “indifference, exhibited by acts of commission or omission, may reach a degree from which an authorization may be inferred.” and that “indifference, exhibited by acts of commission or omission, may reach a degree from which a permission may be inferred” but this does not justify a conclusion that “authorize” and “permit” are synonymous – just as tables and dogs both have four legs, but this does not make tables dogs.

Adelaide Corporation is not authority for identifying “authorize” and “permit”.  Indeed, both it, and Cyril, the case it primarily cites, are authority for distinguishing them, especially given that the legislature chose to use “permit” in section 2(3), and “authorize” in section 1(2).

Second End Note Cyril and iiNet

As an aside here, the Cyril case was concerned with whether the manager of a corporation had authorized the infringement by the corporation’s players of a particular work.  At first instance both the corporation and its manager were found to have authorized.  However, the manager brought an appeal against the decision, and it was his appeal which their Lordships were considering in Cyril.  In many respects Cyril is very similar to the iiNet decision in that the manager may have had the relevant elements set out in section 101(1A) (or at least to a similar level as iiNet has of making available infringements), yet he was still found not to have authorized the performance based on principles of whether, as an officer, he had made the tort his own.

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.

Words in Copyright Act vs Time

Words in Copyright Act vs Time

I have run some numbers on how the size of the (Australian) Copyright Act has changed over the past century or so.  With one exception, these numbers were generated automatically from electronic versions of the legislation.   Before counting the words I stripped out the table of contents and everything from “The Schedule” on.  This is because a bigger Act automatically means a larger table of contents and an older Act means more notes about when sections came into force, were repealed etc.  The one exception is the Copyright Act of 1905, a word count for which was estimated by manually counting words on 3 pages, generating an average per page and multiplying by the number of pages. There are a couple of versions of the Act from between 1905 and the 1970s which are not plotted (as I don’t have access to a full copy of them) but everything I could find from 1970 on is there.

The Statute of Anne (1709) has about 2,500 words in it.  It took roughly 200 years to reach 7,500 (in our 1905 Act).  For your reference, the NSW Conveyancing Act 1919 No6 (which actually does deal with property) has about 82,000 words (ie about the number of words added to the Copyright Act in the last 10 years).  We might speculate whether there will be enough paper in the world to even print the Act in 20 years’ time.

Here is the period from 1992

And, for a laugh, the BSAA reported piracy rate (from the annual reports produced by IDC etc) overlaid.  I would not want to endorse the BSAA numbers as they don’t seem to me to be well founded and any variation in them can easily be explained by changes in IDC’s sampling procedures/calculation methods.  In addition, they are advocacy documents so need to be taken with a grain of salt.

It is hard to see any relationship between the amount of legislation and the reported rate of piracy (which seems to be in a long term, albeit slow, decline).  To the extent there is a relationship, the reported rate seems to lead changes in legislation.

LA Funds Important Legal Research on Free Software Compliance

LA Funds Important Legal Research on Free Software Compliance

Over the past twelve months or so I’ve noticed an upswing in enquiries about free software compliance.  For example, someone might be seeking access to source code for embedded devices with Linux and/or Busybox on them.   One of the key problems for pursuing compliance is the legal concept of  “standing”.   That is, does the court think you have a right to press the claim in question?   So for example, if you see someone (A) breach a contract with someone else (B), a court will probably not let you sue A, basically because that is B’s business [1].  B might not be concerned about the breach, or B might have a relationship with A (or someone else) that might be jeopardised by suing A, so it should be up to B to make the decision about whether to proceed with a suit.  Moreover, A has not infringed a right that you have, so why should you be able to sue?  You’ve not suffered damage, so why should you be able to sue? In short, a court seeks to limit the people bringing actions to only those people whose rights have been infringed.  So,  if you, not holding copyright, see someone breaching the GPL, you can’t sue them in copyright to enforce compliance.

The term “free software compliance” is a short form of compliance with the licence terms for the free software.   In this case, the relevant person whose rights are infringed is the person who granted the licence.  That is, the person who holds the copyright in the software.  Therefore, to bring a court case would require the copyright holders to be parties to the case.  Given that copyright holders are largely concentrated overseas this presents problems for compliance within Australia.  While it is possible for foreign copyright holders to initiate and/or participate in Australian proceedings, it is not simple as the mere distance presents logistical problems.   There are other practical problems that a foreign copyright holder faces.  For example, they may be required to be present in Australia in order to give evidence.   Even where the copyright holder is in Australia they may be reluctant to pursue proceedings.  For example, they may not think it justifies the time commitment they would need to make, or they may not want to be exposed to the possibility of the award of legal costs if they are unsuccessful. I have spoken with some foreign copyright holders and generally they are happy to help, but can’t commit to spend much time helping.

This is where the Trade Practices Act (TPA) is relevant.  The TPA has a very wide concept of standing.  Practically anyone has standing to sue for a breach of the TPA.  If a competitor puts on a misleading advertisement you don’t need to have been misled in order to have it corrected.  You don’t need to show that you (or indeed anyone) suffered a loss.  The mere fact of a breach is usually enough for anyone in the community to enforce compliance with the TPA.   The TPA is therefore particularly relevant for free software compliance.  If a breach of the terms of a licence are also a breach of the TPA, then you don’t need the copyright holders in order to take action.  You can leave them out of it entirely and, instead, rely on your rights as a consumer.

Enforcement under the TPA has other benefits.  First, courts have tended to interpret the Act in a way which favours consumers (not surprising as this is the point of the legislation).  Second, the consumer protections set out in the TPA are pretty straight forward.   You may have heard of “section 52”, which provides that a person must not engage in misleading or deceptive conduct in the course of trade.   In the case of section 52 intention or knowledge is not relevant.  If someone is unintentionally misleading, they will still be caught by the section.   Finally, the TPA comes with an administrative structure for enforcing consumers’ rights in the form of the ACCC and its State and Territory equivalents.  That is, you need not even go to court in order to enforce compliance.  You can, instead, rely on consumer rights procedures (at least in the first instance – if all else fails you may still end up in court,[2] and even then the consumer body may assist you with the case – as happened up to the Federal Court in the groundbreaking Stevens v Sony case).   This may involve some education of the consumer protection people, but, ultimately it’s their job to protect consumers so they must come around sooner or later.  Essentially, vendors who do not (for example) supply source code when they are obliged to are depriving consumers of a legal entitlement.  Imagine how they would react to a vendor selling a car with a wheel missing.

Last year, I was contacted about a potential compliance issue for a router product in which Busybox was embedded (the company in question has since gone into administration).   As  is often the case, the person contacting me did not have a lot of money to spend on legal advice or drafting.   In these cases people don’t tend to contact me until they have first tried to get through to the relevant vendor, so it is usually at at least the first level of escalation.   It seemed that the best way to deal with these cases, at least initially, was to try to create some “self help” materials for people to pursue themselves, without having to engage a lawyer.   To do this, however, would mean that the initial issue of standing needed to be overcome, so I turned to the TPA.  After a short consideration I came to the view that there were profitable lines of inquiry in the TPA, but the mapping of free software to the TPA is not straight forward.  How would the research be  funded?

I was told about the Linux Australia grants system and discussed with one of the then LA committee members whether a “compliance how to” might be worth doing.   I put together two proposals to put some materials together:

Initially, the committee wasn’t convinced that this work would be of value to the community – see after item 5 of this post:

In their next meeting, (last night, 3 March), the Linux Australia committee decided to go ahead and fund the first proposal I submitted (I am making a 1:1 in kind contribution of my time).  This is important research into what arguments might be raised under the TPA if someone is not complying with a free software licence (a decision on the second proposal has been deferred), so the decision is excellent news.  Only yesterday I received another referral – via the SFLC.    They wanted to be able to take it to the next stage and send something more formal to a vendor they’d contacted, but didn’t think it worth spending much money on.  Once the research phase is done, producing notices such as this will be substantially easier (because the background of what is to be said is already set out).   The outputs of the research will be public.

If you know of devices sold in Australia that you think are non compliant, send them to me, because I may as well start a list.

[apologies if you saw a draft of this up – the autosave put it up by mistake]


[1] This is talking only about civil actions.  If there is a crime involved different issues come into play.

[2] The minority of cases end up in court.  By and large things are resolved without recourse to courts.

The Invisible Closed Source Overhead – 3

The Invisible Closed Source Overhead – 3
Brendan Scott January 2009

More on Acquisition –
and About the Day After

Earlier Articles in this Series

The Invisible Closed Source Overhead – 1

The Invisible Closed Source Overhead – 2

How Do You Know What You’re Buying II

Another issue that the acquirer of closed source software may have to contend with is the complexity of the licensing options.  In the previous post we discussed how defining the software which is being acquired can present a problem.  There is an additional, related, difficulty being that the software itself is not necessarily the most important part of an acquisition.  A well defined concept of software in an agreement does not settle the issue because what is being acquired is defined not only by what the software is, but also what can be done with the software as a consequence of the acquisition.   It will be hardly relevant if, even though what constitutes the software has been spelled out in exquisite detail, the licence does not permit the acquirer to (eg) run the software.

Therefore understanding the licensing terms is crucially important to actually defining the acquisition.  Larger vendors can provide a highly differentiated smorgasbord of licensing options, with different things being allowed for different price points and/or configurations.   In some cases whole jobs can be dedicated to simply understanding and explaining the licensing of a single vendor’s suite of products (one example – but job ads tend to be pulled once they expire so google might be a better bet).   As the licence terms effectively define what is being acquired, this means that acquirers can either invest a lot of time and money understanding the range of licensing options – or they can buy blind.  In too many cases it is the latter.

Short Aside

Incidentally, the fact that the same thing is licensed under different conditions depending on its use means that the vendor is effectively charging the customer for the customer’s innovative uses of the vendor’s product.  Imagine if the cost of a pencil differed depending on whether it was being used to write a Nobel-Prize-Winning novel or for a student’s homework – and if you were breaking the law for using the pencil you bought to use at school to later write a Nobel-Prize-Winning novel.

Post Acquisition

Having acquired the software you are now likely to discover that you have a number of additional problems facing you.

Re-engineering Your Processes

Regardless of the process and outcomes of a licence negotiation the vendor will have a pricing model against which you will be charged. In order to support this pricing model you will, at a minimum, need to re-engineer your processes to record any relevant data. Clearly this is not much of an issue for organisation wide one-off up front fee licences. It becomes increasingly complex as the pricing model becomes more fine grained. In addition to pricing requirements some vendors also have copy control requirements as a means of cross checking that the licensing has occurred correctly. This means that manufacturing processes may need to be changed to include, for example, the application of a sticker or other verification to a chassis, as well as processes to order, acquire and store the stock of these indicia.  A vendor may include auditing or accounting requirements, so record keeping practices may also need to be added.  Moreover, there is likely to be wastage of these indicia (lost, damaged etc) so in practice the licence fees for such software will be slightly higher.

Straight Jacketed

Software licence terms are rarely general (eg: “you may install and use the software as is reasonably appropriate”). Rather, they are specific. They will describe in some detail the circumstances in which you can (and can’t) install and use the software. Where you are a manufacturer installing the software on a device you may find the licence includes arbitrary restrictions on the characteristics of the device.

For example  Microsoft recently announced a drop dead date for the shipment of Windows XP.  It then extended that shipment date for devices meeting certain (maximum) specifications.   It subsequently modified that date further, relaxing the limitations on what devices the software could be loaded on [article with overview also here]. So, if you had acted on the earlier indications you would have arbitrarily limited yourself.  The subsequent later easing of the restrictions could end up being to the benefit of competitors (in that you may have committed yourself to a specific platform from which it  takes time to extract yourself).  Indeed (after this paragraph was first drafted) the Vista litigation has shown that exactly this happened.  According to a number of articles (examples: 1, 2, 3, 4)  HP upgraded their stock to match Vista’s requirements, only to find themselves expensive and uncompetitive when the Vista requirements were weakened (at Intel’s insistance).


The negotiation of a closed source licence can be straight forward, or it can be very involved.  The extent to which your business is unusual has an impact on this, with variations from a standard form of operation (as envisaged by the Vendor, albeit based on its other customers) meaning greater difficulty in negotiating a suitable licence.  Ironically, this means that the more innovative your business is, the harder it will be to become licensed.  Equally, if you are in a sector with which the vendor does not normally deal, the licensing terms can be similarly challenging.  Mitch Kapor and Pamela Samuelson, in their lectures at UC Berkeley refer to the case of the university sector.  This sector has very specific requirements for their accounting and calendaring systems.  However, no vendors are prepare to meet those requirements, despite it being a market of substance.  Universities are therefore looking to create open source solutions.

In most cases negotiations lead inexorably to a position which is suitable to each of the parties, although this may take some time.  In some cases though negotiations may break down.  If you have committed time and resources into a product on the assumption that the negotiations will turn out ok, then these will be sunk.  Moreover, you may now be caught with insufficient time to find a replacement and ship your product.  In any case you will end up either curtailing certain aspects of your proposed implementation or paying increased licensing fees or both.

Open source licensors will likely be less flexible on certain points than will closed soruce licensors (who can usually be swayed by the liberal application of licensing fees).  However, it is always clear ahead of time exactly what an open source licensor will be inflexible on, whereas, with a closed source vendor, it may not be until the very end of negotiations that you find out that the vendor really is serious about such-and-such a point.  Moreover, open source licensors will permit broad rights of internal use, so innovative applications of the software are unlikely to cause any licensing issues.  Examples of such problems in the closed source space include multiplexing terminals, embedding on devices and use in virtual machines.

In certain cases (particularly relating to data formats and standards) closed source vendors have an incentive to de-emphasise the licensing terms until after you have become committed to development using the relevant product.  Go have a look at vendor’s websites for their software development kits.  Do these pages emphasise how much fun and productive it is to use the SDK and are covered in eye candy, or do they emphasise the licensing restrictions you’ll be under if you ever want to ship a product using that SDK and are sombre and sober?

What can often happen is that the developers will download and start using an SDK because it solves a problem and legal only gets wind of it when the relevant products are mostly complete and nearing shipment.  This puts the acquirer in a weak negotiating position for the licence terms – or may even tie their hands.

Lack of a Crystal Ball

In many cases the licence you have struck with your vendor will expire after a certain time.  Customers typically enter into such agreements with a poor understanding of how their use of the product will develop over time.  They may believe it is something in the nature of a trial or a stopgap measure.  However, it is not at all unusual for stopgap measures to balloon in scope over time.  At the end of the initial licence period you are then in a position of weakness when renegotiating the licence.  There are a variety of marketing tricks used by Vendors (eg discontinuing old versions or discontinuing support for old versions) to make you pay a new round of licence fees, despite not really needing the difference in functionality.

Further, while this is certainly a problem where software is acquired by way of stopgap, it is also an inherently difficult problem even for those customers with a well developed understanding of their requirements going forward.

Long Term Uncertainty

Open source cannot assure you that you will always have a vendor to turn to when you need to solve a problem.   That’s because a vendor will only exist where there is a community which is sufficiently strong to support that vendor.  What open source does assure you though is that, all else being equal, if there is enough interest then there will be a vendor.   That is, a vendor does not have the option of shutting down a product because they want to focus on a new product.  It can shut down its own support, but it can’t stop others from supporting it – and others will support it if there is a market to do so.

This can happen in a variety of different ways.   A vendor can realise that an existing product effectively cannibalises the market for a new product, so they can stop shipments of the old product.  While you might think that if you stay with your existing copies you’d be fine – which is true to a point.  If you want to exchange data with others, you may be in trouble.  Vendors often introduce file formats in new versions of their products which are incompatible with the formats used in older versions.  If enough people (or enough people in a specific category – government departments are a prime example of this) can be convinced to move to the new format, then, by needing to exchange documents with that category, they can drag existing users forward onto the new format.

The vendor might be acquired by another player who has a different strategic outlook (and therefore cans your product).  The vendor might just lack business acumen and go under.

Developers: Legal Tips for Young Players – Record Keeping

Brendan Scott, October 2008

Records Records Records

In the mid 90s I acted for a large government institution in a dispute over the implementation of a failed search engine development being carried out by one of the worlds’ largest developers/systems integrators.   As is usual in litigation both sides conducted “discovery” – where you produce to the other side any documents of yours which are relevant to the dispute.   Part of the discovery related to the backups of the actual code of the software being developed.  In theory each side should have had a monthly back up for 12 months – that’s what they told us.  When it came to recovering from the backups we discovered most could not be restored.  Out of 24 copies which should have been available, only (I think) 2 could be recovered.   Neither of two large organisations (each of which might fairly be regarded as a paragon of record keeping) were able to keep adequate records.

Small and medium enterprises typically have poor record keeping processes.  Without records, it can be very difficult to prove any point you’re trying to establish – especially if the other side has kept their own (self serving) records.  A good record keeping system will permit you to understand what was happening at a particular time and this can have a lot of important consequences.

Structure Your Records

A record keeping system not only records information, it also structures and stores it in an ordered fashion.   This can sometimes mean having to replicate the record in a number of different places if it happens to be relevant to a couple of different things.   If you don’t keep your records in a structured way you might, in theory, be able to reconstruct what was happening from the jumbled mass of disparate notes you have kept, but, in all likelihood, you won’t be able to do so as a matter of practice.  It is unwise to rely on an after the fact search to find relevant documents.   You might choose to structure things implicitly through the file system.  Alternatively you can use meta-data in a document management system.  However, my experience (now somewhat dated) with document management systems has been that structure is poorly represented (my impression was that they made it easy for me to find other people’s documents, but rarely my own).

Maintain Separation of Projects

A failure to properly structure your files (whether physical or electronic) can have legal implications.  For example, if you fail to maintain a clear demarcation between the open and closed source work that you do, the code from one area may contaminate the other, potentially leading to untold heartache.  Indeed, this applies pairwise to any two  projects you are working on.  Particularly important is that you maintain a strict demarcation between any of your personal coding infrastructure (eg common libraries that you may have developed and reuse for multiple clients) and the specific projects you are working on.

For example, if you were to lump your private libraries into a directory shared with a client project that might seem to a judge (who are not renowned for their understanding of coding practices) as if they formed a single work – which may be bad if you’ve assigned IP as part of your engagement.

Make Notes

In some cases records can serve as a substitute for an express contract.  Equally, where a contract is ambiguous or apparently overreaching, consistent record keeping can, over time, establish some certainty within that ambiguity or can put bounds on the scope of an overbroad contract.  Such records might include, for example, notes of conversations where decisions have been made about what is in or out of scope, or emails which have been sent to this effect.  To the extent there is ambiguity don’t be shy in recording your understanding of the events.  While courts may not necessarily take your view of the world (as instantiated in your records) as gospel, it’s better than having nothing.  Send the notes to the other side prefaced by something along the lines of “This is what I understood we are doing going forward: ….”  This gives them an opportunity to correct you if you’re wrong.  If they don’t it will look odd for them to argue it was wrong at a later date.

Date it

To the extent practicable notes the date the note is taken should be included as part of the note.  This makes it easier to date the note – which can be crucial.  Often cases ride on when a person knew or said something.  If you have an undated note of what was said, it may not end up helping your case if you are unable to place it in relation to some critical date.  For example a note that the other party said “don’t worry, the contract has nothing in it that you should be concerned about” will have a completely  different consequence if it can be proven to be said the day before you sign (perhaps it induced you to sign) as opposed to the day after (when it can have had no effect on your decision to sign).  In addition, courts will generally pay more attention to a record which is contemporaneous with the details it records rather than one which attempts to reconstruct the situation at a later date.

Never use automatic date fields in documents if they will auto-update at some later time.  Date fields may seem like a good idea at the time, but once they update, their value is lost.  If you must use date fields, have some process by which they are converted to plain text on save (eg when the note is finished, or the relevant document has been sent).

Blog Stats

  • 267,289 hits

OSWALD Newsletter

If you would like to receive OSWALD, a weekly open source news digest please send an email to oswald (with the subject "subscribe") at