OPINION OF ADVOCATE GENERAL SIR GORDON SLYNN

My Lords,

Control Data Belgium NV SA (“the applicant”) is a wholly-owned subsidiary of Control Data Corporation, an American company which with its affiliates manufactures computers. The applicant has the exclusive right to import those computers into Belgium for resale or leasing. In 1980 the Université Libre de Bruxelles and the Vrije Universiteit Brussel agreed to hire two computers, known respectively as Control Data-Cyber 170-720 and 170-750, from the applicant for what was considered to be their normal useful life.

An application was made by the universities on 6 August 1980 to the Belgian customs authorities for these two computers (between which no distinction is made for present purposes) to be admitted free of duty pursuant to Council Regulation (EEC) No 1978/75 (Official Journal, L 184 of 15 July 1975, p. 1, as amended by Regulation (EEC) No 1027/79 (Official Journal, L 134 of 31 May 1979, p. 1) and Commission Regulation No 2784/79 (Official Journal, L 318 of 13 December 1979, p. 32). The Belgian authorities were unable to decide whether these computers were scientific instruments or apparatus or whether there existed instruments or apparatus of equivalent scientific value currently manufactured in the Community. They accordingly forwarded the application to the Commission pursuant to Article 7 of Regulation No 2784/79 for a determination whether or not the computers “to be used for research, education and internal administration and in particular in the fields of high-energy physics, plasma physics, thermodynamics, pure mathematics, civil engineering, electrical engineering and nuclear engineering, should be considered as scientific apparatus, and where the reply is in the affirmative, whether apparatus of equivalent scientific value are currently being manufactured in the Community.”

Objections were received to the duty-free importation of these computers from three Member States. Ireland, the Netherlands and the United Kingdom, all on the basis that these were not scientific instruments or apparatus, and the last-mentioned additionally on the basis that an equivalent computer could be obtained in the Community. The Commission accordingly referred the question to a committee of experts acting within the framework of the Committee on Duty-Free Arrangements, which at a meeting in June 1981 decided that these computers did not have “a scientific character”.

The Commission on 10 August 1981, in a decision addressed to the Member States and numbered 81/692/EEC concluded that the computers did not have “the requisite objective characteristics making them specifically suited to scientific research”, that “apparatus of the same kind are principally used for nonscientific activities” and that their use by the universities could not confer on them the character of scientific apparatus so that they could not be imported duty-free as scientific apparatus.

By an application dated 20 November 1981 under Article 173 of the EEC Treaty, the applicant asked the Court to annul the Commission's decision.

The legislative provisions can be summarized shortly:

(a)Regulation No 1798/75, having recited that, in order to facilitate the free exchange of ideas and scientific research, it is necessary to allow “by all possible means” the duty-free admission of educational and scientific materials, and that the Florence Agreement, drawn up in 1952 under the auspices of Unesco, which was in force in the majority of Member States, required such duty-free admission to be granted, provided by Article 3, that in addition to specified items which were to be imported duty-free, “scientific instruments and apparatus” not so specified which were “imported exclusively for educational purposes or for pure scientific research may be admitted duty-free by the establishments described (which it is agreed include the two universities) so long as items of equivalent scientific value were not being manufactured in the Community.

(b)On 26 November 1976 a Protocol to the Florence Agreement was adopted in Nairobi with the purpose of making it easier to import educational and scientific materials. This replaced the requirement that apparatus should be admitted duty-free only if it was intended “exclusively” for educational purposes or “pure” scientific research, by a provision that scientific apparatus should be so admitted unless it was intended for commercial purposes.

(c)Council Regulation No 1027/79, recognizing that the Protocol intended to extend relief from duties to products up to that time excluded from such relief and that Community legislation should be amended to take account of the provisions of the protocol to the extent compatible with the Community's aims, introduced an important amendment to Article 3 of Regulation No 1798/75. From 1 January 1980, unspecified scientific instruments and apparatus intended for establishments (which include the two universities) were to be admitted duty-free if they were “imported exclusively for noncommercial purposes”. For the purposes of the article “a scientific instrument or apparatus” shall mean “any instrument or apparatus which, by reason of its objective technical characteristics and the results which it makes it possible to obtain is mainly or exclusively suited to scientific activities”.

(d)Article 5 of Commission Regulation (EEC) No 2784/79 took the matter further. The “objective technical characteristics” of a scientific instrument or apparatus “shall be understood to mean those characteristics resulting from the construction of that instrument or apparatus or from adjustments to a standard instrument or apparatus which make it possible to obtain high-level performances above those normally required for industrial or commercial use”. If it is not possible to establish clearly on the basis of its objective technical characteristics that an instrument is a scientific instrument “reference shall be made to the general uses in the Community” of the particular instrument. If it is shown that the instrument is used “mainly” for scientific purposes, it shall be deemed to be of a scientific nature.

The Court's reference in the Utrecht case to the objective characteristics of an instrument must, thus, now be read subject to the definition of “objective technical characteristics” which is contained in this last regulation.

The Commission takes the preliminary objection that this application is inadmissible.

In the first place it is said that the decision is not of individual concern to the applicant, even if of direct concern, within the meaning of Article 173 of the Treaty. The decision covers all equipment of the same type which is imported into the Community and must be applied by all Member States. It should therefore be regarded as a decision in rem rather than in personam. Moreover the applicant is not a member of a fixed and identified class as defined by the Court in a series of cases. It is merely one company in a group whose parent makes and controls the marketing of these computers in the Community. Others may be given the right to import the computers during the validity of the Commission's decision and even other subsidiaries of the parent company may come into existence and be given the right to import. It is only the parent company which is both directly and individually concerned. Accordingly the proper course here would have been to attack the decision by proceedings in the Belgian courts, which was not done.

I do not accept these contentions. There are different types of computer designed for different purposes and a decision on a particular model may be distinguishable when some other manufacturer's product is considered. The case is wholly different from those where the sellers and buyers of agricultural products are concerned. It may well be that the applicant is only one member of the group and that the latter operates on a common basis controlled by the parent company. There is nothing to suggest the likelihood of outsiders being licensed to import the products, and it seems to me too technical and legalistic to rely on the separate corporate identity of the applicant in the present case. In my opinion this company is just as directly and individually concerned as the parent company. There are here attributes peculiar to the company, and circumstances in which it is differentiated from other persons, within the meaning of such cases as Plaumann v Commission [1963] ECR95 at p. 107.

Secondly it is said in the defence that this is not a decision which represents the culmination of the procedure so far as the applicant is concerned. It is an act internal to the Community and the Member States: the final act is that of the Belgian authorities who implement the decision. This point was said at the hearing not to be “insisted” upon. I did not understand that this amounted to an abandonment of the point but only to a realistic recognition of its lack of force. Clearly this was the culmination of the procedure so far as the Commission was concerned, and no further decision by the Commission on this application fell to be taken. It is taking the Court's decisions on this requirement too far, to argue that the fact that a national authority's decision has to follow prevents the Commission's decision from being a final decision, not least when the national authority automatically applies the Commission's decision. Whether or not it would have been better for this challenge to have been brought against the Belgian authority's implementation of the Commission's decision in the national court, it seems to me that there is here a decision by the Commission within the meaning of Article 173.

Accordingly in my opinion this application is admissible.

The applicant contends that there are defects of procedure and form which vitiate the decision.

First, it is said that the decision does not adequately state reasons contrary to Article 190 of the Treaty. It repeats, parrot-like, the phraseology used in a number of other cases where computers have been denied duty-free status, in none of which have adequate reasons been given. As a result, there is no material upon which the applicant and the Court can be satisfied that the Commission has proceeded upon a proper basis, or acceptable reasoning. None of the detailed arguments have been answered: none of the factual material contained in a 50-page submission to the Belgian authorities is the subject of comment.

Secondly the applicant contends that there has been no or no adequate opportunity to address the Commission, to give explanations, to answer points taken against it, to know what was in the Commission's mind. The procedure adopted contrasts so vividly with the procedure followed where violations of Articles 85 and 86 of the Treaty are involved that it must be declared to deprive the applicant of its essential procedural rights and safeguards. Even if, as is accepted, the Commission has followed the strict letter of the procedure laid down in the regulations, fairness required in the particular circumstances of this case that the applicant should be told of the case against it and be allowed to reply.

I do not consider that the decision is given in a satisfactory way. This is an important case, carefully prepared and fully argued on paper for the purposes of the application to the Belgian authorities. Substantial sums of duty are at stake. The amendments introduced by the 1979 regulations are relatively new and the Commission has not in the decision spelled out its attitude and indicated the criteria it adopts, let alone answered the particular features of these two computers which are put forward. It would have been better, in my view, if it had done so. Nonetheless, it seems to me that it is possible to glean from the stark words used the basic reasons behind the decision. Whether the Commission is right or wrong, the reasons are (a) that the objective characteristics (those set out in Article 5 of Regulation No 2784/79) are not shown to exist; (b) as a fact, computers of the same kind are principally used for nonscientific activities; (c) the fact that they may be used here for scientific purposes is not enough to deem them to be scientific apparatus.

As to the second argument, it is clearly wrong to argue, and at the end of the day the applicant made it clear that it did not argue, that because a better procedure is adopted for other types of case, the procedure contemplated by these regulations is defective in such a way that the Court can interfere. On the facts of the particular case, I do not think that the applicant is entitled to an oral hearing (nor as I understand it did it ask for one) or that it can say that it did not have a chance to put its own case forward. The very submission which is relied on as calling for detailed reasoning in the decision contains in full the arguments in favour of the applicant's case. Moreover, even if the applicant did not know, from the terse wording of earlier decisions, the precise reasons why other computers had been rejected, it appears likely at least to have been aware that the Commission did not as a broad policy accede to the duty-free importation of computers unless they were incorporated in other instruments which taken as a whole were regarded as scientific in their own right.

There are instances where decisions of this kind have to be taken quickly, and where the procedure must be tailored accordingly. This was not such a case, and it would have been more satisfactory if the applicant had been told of the answers to the specific points made by it. Nonetheless in the end it is clear that the Commission would have resorted to its “broad-brush” approach, and I am not satisfied that the applicant has been prejudiced by the fact that it was not told of the individual objections to specific points raised.

Accordingly, and although I regard this as an unsatisfactory case as far as the first ground is concerned, I do not consider that it would be right to annul the Commission's decision on either of these grounds.

The applicant's principal claim is that these computers are scientific instruments or apparatus, and that the Commission has come to a wrong decision on the facts and has misdirected itself in its approach when deciding that they were not. Whereas other manufacturers produce computers for commercial or industrial use, the Control Data Group has concentrated on computers specifically designed for high-performance application in scientific research. The “hardware”, the basic structure of the computer, as well as the software, are geared to work at high speed and with great precision. The hardware prepared for these scientific purposes cannot be changed or can only be changed with great difficulty. Whereas the computer used for commercial purposes is principally required to take in instructions from the operator, to record them and to move data about within the computer, and to carry out relatively simple mathematical calculations, the computers in question are concerned with highly complex mathematical calculations involving little movement of word data. Very large or very small numbers can be used which would have no relevance to commercial uses. The units, groups of electronic switches found in the Cyber computers, are larger so that longer numbers can be used. These are needed for mathematical or scientific calculations but the units are wasted on the smaller numbers used for commercial transactions. Moreover, there is a significant difference between the two in that the central processing unit of the computers is connected only with the memory of the machine and instructions cannot be fed in directly from external sources which on each instruction would otherwise interrupt the activity of the computer. A different programming language is used for scientific work than for commercial work which is word-rather than number-dominated. Evidence has been given of the kind of research which can be done in studying the behaviour of electrons, in plasma and particle physics and in weather forecasting. The speed of the computers is such that for example the solving of equations which would take years for the individual scientist can be done in minutes or hours. The behaviour of, for example, atomic particles can be calculated or predicted mathematically in a way which would be difficult or impossible by physical experiment. Without the ability to make calculations and comparisons at high speed it is said that much contemporary scientific research would be, if not impossible, at least long delayed. Even though administrative work can be done on the Cyber computer (and is done here for 5 or 7% of the time) it is cumbersome and inefficient and is only done on these computers because the universities have them and because it would be unjustified to have a commercial-type computer in addition merely for that other type of work, even though a commercial-type computer would be more efficient.

On the facts, the Commission accepts that the computers are capable of complex calculations and well-adapted to scientific calculation. The greater the computer's capacity, the faster the computer can carry out complex calculations and it is acknowledged that these are amongst the largest computers available. On the other hand the Commission challenges the assertions that word-orientated hardware and floating point arithmetic, relied on by the applicant, are specific to the Cyber computer, and says that they are commonly found in computers. What really determines the scientific use of a computer is the software, and this shows that the hardware of the computer is not scientific apparatus. These computers can use both the number-dominated (Fortran, formula translation) language systems and the word-dominated systems (Cobol, common business oriented language). Moreover, even with the Fortran system, purely administrative tasks can be performed, so that these should be regarded merely as “an independent calculating tool capable of operating at high speed, designed, for a wide range of types of calculation and administration applications and not for any specific form of scientific research”.

So far as the interpretation of the regulations is concerned, the Commission takes as “an extremely important point” that a computer is not “an instrument”, since it is impossible to measure, reveal, transform or process any physical dimension or characteristic with a computer. It is a “technical tool on a par with, for example, an adding machine”. A scientific instrument must be able to deal with physical properties with a view to carrying out experiments, and computers only operate on numbers, not on physical properties.

I find this argument impossible to accept. I do not read any such limitation into the ordinary meaning of the word or see that it is to be implied from anything to be found in the regulations. Equipment capable of doing complex mathematical equations is in my opinion “an instrument”; if it is not it is in any event “apparatus”.

Then, it is said, as the Commission's general and fundamental contention, that computers are not scientific instruments. This does not mean merely that computers like those in question here are not scientific instruments. Computers as such are not scientific instruments. As I understand the Commission's contention, it is that computers as such are not capable, as a matter of construction of the regulation, of being scientific instruments. Scientific instruments must be able to deal with physical properties with a view to carrying out experiments. Computers do not operate on physical properties but only on numbers. Therefore they cannot, apparently, be scientific instruments. The Commission “does not deny the contribution of computers to scientific research” but says that the process of calculation is merely auxiliary in the scientific field. It is the Commission's unvarying practice to allow the admission duty-free of computers only when they are incorporated in other equipment which, taken as a whole, is regarded as a scientific instrument in its own right.

That submission, again, I consider to be untenable. This case is not concerned with what “scientific instruments” may mean in general parlance or in any other context than that of these regulations. Regulation No 1027/79 makes it plain that any instrument or apparatus, seen by the light of its objective characteristics and the results it makes it possible to obtain, which is mainly or exclusively suited to scientific activities, is a scientific instrument or apparatus for the purpose of Article 3 of that regulation. It does not seem to me possible to say that equipment which does extremely complex mathematical equations and calculations at high speed is, as a matter of law, incapable of being a scientific instrument, any more than it would be possible to say that because a child may use a simple microscope to look at natural objects as a hobby, therefore microscopes cannot be scientific instruments. What must be asked in each case, once it is accepted that “the unit” is an instrument or apparatus, is whether the other parts of the definition in Article 3 of Regulation No 1027/79 and Article 5 of Regulation No 2784/79 are satisfied.

The Commission contends that because of the fact that the computers can be used for, in addition to research, teaching and internal administrative tasks, they are to be regarded as “all-purpose” equipment rather than scientific instruments. It seems to me to be an infelicitous use of language to describe as an all-purpose or general-purpose instrument one which is designed and intended for one use, particularly suited to that use and principally put to that use, merely because, since it is there, it is used albeit inefficiently for another purpose for at the most 7% of the time. Be that as it may, it does not seem to me to prevent a computer from being a scientific instrument within the meaning of the regulation merely because it is capable of being used, and is used to a limited extent, for other purposes. Such a conclusion would in any event be in conflict with what was decided by the Court in the Utrecht case as set out earlier in this opinion.

Although “scientific” and “commercial” uses do not seem to me as a matter of language to cover the whole field of uses to which a computer may be put, it is to be borne in mind that a scientific instrument to qualify must be:

(a)imported exclusively for noncommercial purposes;

(b)intended for establishments “principally” engaged in education or scientific research;

an instrument which,

(i)by reason of its objective technical characteristics, namely those characteristics resulting from the construction of that instrument or apparatus which make it possible to obtain high-level performances above those normally required for industrial or commercial use, and

(ii)the results which it makes it possible to obtain

is mainly or exclusively suited to scientific activities.

I do not understand it to be challenged that these computers were imported exclusively for noncommercial purposes; if it is, then I would reject the challenge. Since the universities are authorized, apparently, to receive scientific instruments duty-free, and for other reasons, it is also clear that the universities are accepted to be “principally” engaged in education or scientific research. Paragraphs (a) and (b) above are therefore satisfied.

In view of the fact that the Commission accepts that computers are capable of complex calculations and are “well-adapted to scientific calculations” and in view of the unchallenged evidence as to the use to which these computers can be put, and are put, and to which some other types of computer cannot be put, it seems to me that it is clearly shown here that it is “possible” (not “only possible”) to obtain high-level performances above those normally required for industrial or commercial use.

The remaining question is whether because of these characteristics, and the results which they make it possible to obtain, these computers can be said to be “mainly” suited to scientific activities. “Mainly suited to” does not mean “only usable for”; and “scientific activities” is wider, in my view, than “scientific research”. I leave aside the word “exclusively” despite the argument that something can be suited exclusively to one purpose even if it can be used “unsuitably” or inappropriately for another, since if the computer is not “mainly suited to” scientific activities, it cannot be “exclusively suited to” them.

On the evidence before the Court it does not on any view seem to me to be possible to conclude that (c) above is not satisfied, in other words that on the facts the Commission came to a right conclusion.

In the ordinary way, if the factual evidence were conflicting and difficult to resolve, so long as the Commission could be shown to have directed itself properly as to the proper construction of the regulations and as to the correct approach, I would think it wrong to interfere with the decision of the Commission. In my view it would not be right to take that course in the present case. The Commission has erred in law as to the right approach (as to the meaning of “instrument”, whether computers can be scientific instruments and as to the effect of a machine being capable of use for other limited purposes) and the factual evidence on behalf of the Applicant has not been displaced.

I consider that this decision should be annulled.

If “objective technical characteristics” had not been defined in the way it has, I should have considered it necessary to annul the decision without expressing a conclusion as to whether these computers were scientific instruments, on the basis that this exercise could only be performed by the Commission on an assessment of expert evidence. That is not the position. There is a definition, and it seems to be established clearly here that the characteristics (e. g. speed and the complexity of the calculations which are possible) resulting from the construction of the instrument (e. g. the link of the central processing unit with the memory and the inability to make “interrupts” by tapping in instructions externally, the word-orientation and word length used, the sacrifice of part of the machine's capacity in order to use Fortran and the other characteristics referred to by Dr Jackson and in the written evidence and pleadings) are such as to achieve high-level performances above those normally required for industrial or commercial use.

Although the Commission rejects some of the contentions put forward by the applicant, it seems to me that in substance it does accept that the size and construction of these computers enable complex calculations to be done at high speed in a way eminently suitable for scientific research, and which it is not said are necessary normally for industrial or commercial use. Accordingly it seems to me that (a), (b) and (c) have been satisfied, and that the Commission erred both as to its interpretation and its application of the regulations and thereby in the use of the powers entrusted to it in rejecting the applicant's claim.

If I had not been so satisfied, I would have been of the opinion that the decision should be annulled on the basis that the Commission has incorrectly approached this decision and the Commission would then need to reconsider the application in the light of the considerations to which I have referred.

I have not dealt with the second paragraph of Article 5 (1) of Regulation No 2784/79 because there is a conflict of evidence as to the uses to which other computers of these models are put. If the decision were annulled on the second basis referred to above, it would be for the Commission to decide in the light of the considerations to which I have referred, whether the second paragraph is satisfied should the Commission not be able to establish clearly on the basis of their objective characteristics (as defined) whether these computers were to be regarded as scientific instruments or apparatus.

I have deliberately left until last the fact that the Commission has referred to a working document dated 12 February 1980 prepared by the Committee on Duty-Free Arrangements for the purpose of applying Regulation No 1798/75 as amended and Regulation No 2734/79. Whilst recognizing the value of working documents such as these for those who apply or who are affected by the regulations, it does not seem to me right to have regard to this document in construing the regulations. It does not, in any event, inhibit the conclusions expressed above. To summarize the notes:

Normally computers normally used for production, commercial applications of production, routine analyses or other nonscientific processes are outside the scope of the description of scientific instruments. If anything, this suggests that computers can be scientific instruments if they are used as part of scientific processes.

Instruments excluded by paragraph 1.3.1 may “in certain circumstances” be regarded as scientific instruments if there have been “additions or substantial modifications which have the effect of rendering them specifically suitable for research or educational purposes”. Those modifications or additions could, in my view, be made in the course of initial manufacture, as well as subsequently, if changes were made to an otherwise standard and non-qualifying model. If, as I consider, a computer is an instrument, there can be no possible justification for allowing in a modified standard version, but excluding a model which is from the beginning “specifically suitable” for research or educational purposes. “Specifically suitable for” does not mean “exclusively usable for”, and an instrument can be specifically suitable for one purpose even if capable of use for another.

Finally, an instrument may be admitted if it is incorporated into a unit which taken as a whole is considered scientific.

I do not consider that the specific exemptions in paragraphs 1.3.2 and 1.3.3 in any event give the exclusive categories of exceptions to the normal rule.

In the light of these matters it is my opinion that Commission Decision 81/692/EEC of 10 August 1981 should be annulled.