The Court of Appeal has dismissed the SAS appeal against the ruling of Arnold J in the High Court. But the comments on the High Court judgment and the reasoning behind it are of real interest.
In SAS Institute Inc v World Programming Ltd  EWCA Civ 1482, the Court of Appeal had to give judgment on the extent to which the developer of a computer program may lawfully replicate the functions of an existing computer program; and the materials that he may lawfully use for that purpose.
In the High Court, SAS alleged that in creating its World Programming System, WPL had:
· used the SAS Manuals as a technical specification for WPS and copied a substantial part of those manuals in its system, thereby infringing copyright in the SAS Manuals (the 'Manual to Program Claim');
· indirectly infringed copyright in the SAS Components in creating its system (the 'Program to Program Claim');
· infringed the copyright in the SAS Manuals by reproducing a substantial part of them (the 'Manual to Manual Claim'); and
· repeatedly used the SAS Learning Edition outside the scope of the applicable, thereby infringing the SAS copyright and acting in breach of contract (the 'Learning Edition Claim').
The proceedings followed a torturous course that required a reference to the ECJ, but in his third judgment (after a valiant attempt to fathom the meaning of the answers of the ECJ), Arnold J dismissed the SAS claims, except insofar as he found limited breaches of copyright in relation to the Manual to Manual Claim. He gave permission to appeal on the Learning Edition Claim, and the Court of Appeal gave permission to appeal on the Manual to Program Claim, and the remainder of the Manual to Manual Claim. There was no appeal against the dismissal of the Program to Program Claim as, following the ECJ judgment, it was common ground that neither the SAS Language nor the functionality of the SAS System is protected by copyright under the Software Directive.
In a detailed judgment, Lewison LJ initially focused on the creation of the WPL program by reference to the SAS manuals. In essence it was alleged that WPL in writing its system in Java (and subsequently in C++) had copied the expression of the intellectual creation of the author of the SAS Manuals. Lewison LJ observes that 'this argument seems at first sight to be counter-intuitive, because the SAS Manuals themselves do not contain any programming language'. After considering various cases and the ECJ responses to the questions referred to it in this dispute, Lewison LJ states that, while Arnold J may have been partially misled in his reading of the ECJ judgment ('the judge does not seem to have appreciated that the judgment of the CJEU had, to some extent, changed the question'), he nevertheless reached the right conclusion. Lewison LJ agreed 'with the Advocate-General's pithy conclusion that "The WPL System does not reproduce the description [in the SAS Manuals] of those statistical operations but simply executes them."'
Lewison LJ did highlight one important point, distinguishing his view from that of Arnold J. Where it was suggested by Arnold J that the SAS System was not the intellectual creation of its authors because it had grown by accretion without an overall design, Lewison LJ felt that that was not a reason to deny it protection.
Subject to a procedural point (which received no welcome from Lewison LJ), the appeal in relation to the Manual to Manual Claim, or what remained of it on appeal, was said to stand or fall with the Manual to Program Claim. So that element of the appeal also failed.
Readers may find all the comments on WPL's alleged breach of the SAS Learning Edition licence of considerable interest (from ) but, in dealing with the SAS complaint that WPL had exceeded the contemplated use of the Learning Edition, Lewison LJ may surprise some with these observations which may be of relevance in any number of software licensing contexts:
Finally, it is worth noting that Lewison LJ was not greatly impressed by the contribution of the ECJ to the resolution of this dispute, saying (at ):
Unfortunately the parties could not agree what the CJEU had actually decided. The language in which the court expressed its judgment was, at times, disappointingly compressed, if not obscure. Moreover, although the judge had referred specific and detailed questions to the CJEU, the CJEU refrained from answering them, but instead answered its own paraphrase. This led to a disagreement about whether the court had actually given answers to all the questions posed. It would, perhaps, be more helpful if in response to a national court asking for help the CJEU, in the performance of its duty of sincere co-operation, answered the questions it was asked unless there are cogent reasons not to.