---

REPUBLIC OF SOUTH AFRICA

IN THE HIGH COURT OF SOUTH AFRICA

(WESTERN CAPE DIVISION, CAPE TOWN)

Case No: A504/13

DATE: 09 OCTOBER 2014

REPORTABLE JUDGMENT

In the matter between:

THE TRUSTEES FOR THE TIME BEING......................................First appellant

OF THE DELSHERAY TRUST

ELSA JOHANN BOTHA............................................................Second appellant

DESMOND BOTHA.....................................................................Third appellant

And

ABSA BANK LIMITED....................................................................Respondent

JUDGMENT DELIVERED ON 9 OCTOBER 2014

BLIGNAULT and HENNEY JJ

[1] This is an appeal from an order of a single judge of this division in terms of which summary judgment was granted in favour of respondent against appellants. The issue on appeal concerns the admissibility and adequacy of respondent’s verifying affidavit which is based exclusively on its computerised records.

Respondent’s particulars of claim

[2] Respondent instituted the action against first appellant, the Trustees for the time being of the Delsheray Trust, as principal debtor and second and third appellants as sureties in respect of first appellant’s debt. Respondent claimed payment from appellants, jointly and severally, of the amount of R1 588 208,85 plus interest thereon accrued at a rate of 8,5% calculated daily and capitalised monthly from 19 September 2012 to the final date of payment.

[3] Respondent’s cause of action is a breach of an agreement of loan between it and first appellant in terms of which it lent and advanced three separate amounts totalling R1 700 000,00 to first appellant. The first advance was in the amount of R900 000,00, the second in the amount of R200 000,00 and the third in the amount of R600 000,00.

[4] As security for the repayment of the loans respondent registered three covering bonds over first appellant’s immovable property described as Erf 4805 Eversdale.

[5] According to respondent’s particulars of claim the original contract documents were destroyed in a fire. In an affidavit deposed to on behalf of respondent and attached to the particulars of claim it was alleged that the details of first appellant’s account reflected in respondent’s particulars of claim were extracted from its computerised records.

[6] Respondent alleged that second and third appellants bound themselves as sureties and co-principal debtors with first appellant for the due repayment of the monies lent and advanced by respondent to first appellant.

[7] In addition to the claim for payment of the amount of R1 588 208,85 respondent sought an order declaring the mortgaged property specially executable in terms of the provisions of the mortgage bonds.

[8] In response to the summons, appellants filed a notice of intention to defend the action, whereupon respondent applied for summary judgment.

The affidavits

[9] Respondent’s affidavit in support of its application for summary judgment was deposed to by Mr Yuven Pillay. It reads as follows:

‘1) I am a Specialist employed at the Retail Bank Collection Division of the Plaintiff / Applicant. I am duly authorised to depose of this Affidavit. All the data and records, relating to the Applicant’s/Plaintiff’s action against the Defendant (Appellants) are under my control and I deal with this account on a day to day basis. The facts contained herein are within my personal knowledge and are both true and correct.

2) Unless the contrary appears, I have knowledge of the facts hereinafter stated, either personally or as a result of my access to all relevant computer data and documents pertaining to the Trust’s mortgage loans, account number 4056939083.

3) I hereby verify the facts and cause of action stated in the Summons and the Particulars of Claim to the Summons as true and correct and verify in particular, that the Respondents/Defendants jointly and severally the one to pay the other to be absolved are indebted to the Plaintiff in the sum of R1 588 208,61 on the grounds stated in the Summons.

4) In my opinion the Respondents/Defendants, jointly and severally, the one paying the other to be absolved, do not have a bona fide defence to the Applicant’s/Plaintiff’s claim and their appearance to defend has been entered solely for purpose of delay.’

[10] First appellant deposed to an answering affidavit on behalf of all three appellants. His principal defence, raised in limine, was that the verifying affidavit made by Mr Pillay did not comply with the provisions of Rule 32(2) of the Uniform Rules of Court (‘Rule 32(2)’). The substance of this defence is that Mr Pillay purported to rely on the computer data and records of respondent and not on his own direct personal knowledge. First appellant did not in his answering affidavit deal with the merits of respondent’s claim against appellants at all.

[11] The court a quo granted summary judgment as prayed against the first, second and third appellants.

[12] With the leave of the court a quo, appellants appealed against the summary judgment granted against them. Respondent has, however, abandoned its judgment against the second and third appellants and this appeal concerns only the summary judgment granted against the first appellant.

The central issue

[13] In terms of Rule 32(2) a plaintiff seeking an order of summary judgment is required to file an affidavit in support of the application made ‘by himself or by any person who can swear positively to the facts verifying the cause of action and the amount claimed and stating that in his opinion there is no bona fide defence to the action ….’ I shall refer in this judgment to such affidavit as a verifying affidavit.

[14] The verifying affidavit of Mr Pillay is not a model of clarity. It is clear that he relied at least to a significant extent on the computerised records of respondent. He did not state, however, that he relied on specific facts dehors respondent’s computerised records. He did not state, for example, that he had any direct personal knowledge of the conclusion of the agreement of loan between first appellant and respondent or the terms thereof or any of the transactions reflected as debits and credits on first appellant’s account. We shall accordingly approach the issue in this case on the assumption, in favour of first appellant, that Mr Pillay relied exclusively on respondent’s computerised records, ie he purported to swear positively to the facts revealed to him by respondent’s computerised records.

[15] Counsel for first appellant argued that Mr Pillay did not state that he had consulted with any witnesses in order to gain personal knowledge of the facts of the matter. By relying exclusively on respondent’s computerised records he did not purport to have any direct first-hand knowledge of respondent’s cause of action or the quantum of its claim. The verifying affidavit, according to the argument, did therefore not comply with the provisions of Rule 32(2). Counsel also argued that the fact that the original contract documents were destroyed, strengthened his submission as Mr Pillay would have been unable to acquire personal knowledge thereof.

[16] The central issue in this case is therefore whether Mr Pillay’s verifying affidavit which is founded exclusively on respondent’s computerised records, complies with the provisions of Rule 32(2). For the reasons that follow we are of the view that it does.

[17] The subject of computer evidence in South Africa, it should be noted, is regulated by the Electronic Communications and Transactions Act 25 of 2002 (‘ECTA’) but respondent did not present or argue its case on the basis of the provisions of ECTA. In terms of the express wording of s 3 of ECTA, however, it does not exclude the application of the common law. We shall accordingly deal with the admissibility and adequacy of Mr Pillay’s verifying affidavit in terms of common law principles.

The information revolution

[18] It is well known that modern technological developments have brought about a revolution in the way that information, including legal information, is captured and disseminated. These developments brought about substantial changes in the law of computer generated evidence, internationally and in South Africa. Although well known, a few quotes may not be out of place.

[19] As long ago as 1989 Steyn J said this in R v Minors R v Harper [1989] 2 All ER 208 (Court of Appeal, Criminal Division) at 210:

‘The Law of Evidence must be adapted to the realities of contemporary business practise. Main frame computers, mini computers play a pervasive role. Often the only record of a transaction which nobody can be expected to remember, will be in the memory of a computer. In versatility, power and frequency of use of computers will increase. If computer output cannot relatively readily be used as evidence in criminal cases, much crime and notably offences of dishonesty will in practice be immune from prosecution.’

[20] Mason Electronic Evidence 3rd edition (2012) para [1.34] says the following:

‘Now we live in the age of the machine, the range of digital evidence that is capable of being captured, investigated and disclosed in legal proceedings is very wide. From the files on a digital camera to the complex behaviour of a computer attached to the Internet, assessing digital evidence has become the staple of a lawyer's life.’

[21] The remarks of George L Paul in a review of Mason’s Electronic Evidence supra, published in 53 Jurimetrics – (Summer 2013) 467, are equally illustrative:

‘…. there is nothing more important to a legal system than the analysis of “information.” Information is the lifeblood of commerce, societal dialogue, and interpersonal communication; and therefore, it is also the subject of litigation the world over. Our economy, indeed, is now overwhelmingly an “information economy,” and will continue to become increasingly so in the future.’

South African case law

[22] The question of the adequacy of the verifying affidavit in summary judgment proceedings has been considered in various recent judgments. All of them refer back to the locus classicus, the judgment of Corbett JA in Maharaj v Barclays National Bank Ltd 1976 (1) SA 418 (A). The following two oft quoted passages in that judgment have a bearing on the present enquiry:

At 423 BC:

‘Generally speaking, before a person can swear positively to facts in legal proceedings they must be within his personal knowledge. For this reason the practice has been adopted, both in regard to the present Rule 32 and in regard to some of its provincial predecessors (and the similar rule in the magistrates’ courts), of requiring hat a deponent to an affidavit in support of summary judgment, other than personal knowledge or make some averment to that effect), unless such direct knowledge appears from other facts stated’.

And at 423 H – 424 F:

‘Ex facie the summons plaintiff's cause of action is founded upon moneys disbursed on defendant's behalf in terms of an oral agreement of overdraft. The relevant facts would, therefore, be the conclusion of the contract, and the terms thereof, the deposits in, and withdrawals from, defendant's current account at the Stanger branch of the plaintiff bank and the interest debits resulting in the debit balance as at the date alleged in the summons, viz. 24 October 1974, and the making of a demand for payment. In regard to certain of these facts, it would be difficult, if not impossible, for any one person to have first-hand knowledge of every fact that goes to make up the plaintiff's cause of action. In this connection I am in full agreement with the following remarks of MILLER, J., in Barclays National Bank Ltd. v. Love, supra at pp. 516 - 7, made with reference to an affidavit made by the manager of a branch of the plaintiff bank (oddly enough also the Stanger branch):

“We are concerned here with an affidavit made by the manager of the very branch of the bank at which overdraft facilities were enjoyed by the defendant. The nature of the deponent's office in itself suggests very strongly that he would in the ordinary course of his duties acquire personal knowledge of the defendant's financial standing with the bank. This is not to suggest that he would have personal knowledge of every withdrawal of money made by the defendant or that he personally would have made every entry in the bank's ledgers or statements of account; indeed, if that were the degree of personal knowledge required it is difficult to conceive of circumstances in which a bank could ever obtain summary judgment. It goes without saying that a manager of a bank who claims to have personal knowledge of the extent to which a client has overdrawn his account must needs rely upon the bank records which show the amounts paid into his account and the amounts withdrawn by the client.”

In this case the deponent, Mr Mason, does not specifically state that he has personal knowledge of the overdraft arrangements made by the defendant with the manager of the Stanger branch of the bank and the state of defendant's current account at the relative time. On the other hand, he does say, in para. 1 of his affidavit, that he is the assistant to the branch manager of the Stanger branch. It is not clear what the duties or status of the assistant are but, if one reads this averment together with the statement in para. 2 that the deponent swears positively that the defendant is liable to plaintiff on the claim and for the amount as detailed in the summons and upon the cause of action as set out therein, there is perhaps enough to justify the conclusion that in the course of his duties Mr Mason would have acquired a personal knowledge of the defendant's financial standing with the bank and the state of his current account.’

[23] There is a line of judgments of the high courts which are to the effect that the deponent of a verifying affidavit in summary judgment proceedings cannot rely exclusively on a perusal of records and documents of the plaintiff for purposes of that affidavit. The judgment of Wallis J in Shackleton Credit Management (Pty) Ltd v Microzone Trading 88 CC and Another 2010 (5) SA 112 (KZP) is typical of this approach. See para [13] of the report:

‘[13] It may be … … that first-hand knowledge of every fact which goes to make up the applicant's cause of action is not required, and that where the applicant is a corporate entity, the deponent may well legitimately rely on records in the company's possession for their personal knowledge of at least certain of the relevant facts and the ability to swear positively to such facts. However, I do not understand any of the cases as going so far as to say that the deponent to an affidavit in support of an application for summary judgment can have no personal knowledge whatsoever of the facts giving rise to the claim, and rely exclusively on the perusal of records and documents in order to verify the cause of action and the facts giving rise to it.’

[24] The same approach was approved and applied by Southwood J in Standard Bank of South Africa v Han-Rit Boerdery CC and Others [2011] ZAGPPHC 120 (22 July 2011) para [6]:

‘In both the Shackleton case (paras 7 and 13) and the Beyer case (paras 9, 10, 19, 20 and 21) the court found that a deponent who acquires his knowledge from documents to which he has access cannot swear positively to the facts. In both cases the courts reviewed the relevant case law and the principles laid down over the years and I respectfully agree with the reasoning of the courts and the conclusion reached.’

[25] The judgment of Binns-Ward J in Absa Bank Ltd v Le Roux and Others 2014 (1) SA 475 (WCC) is to the same effect. He followed the Shackleton judgment and arrived at a similar result. He pointed out, inter alia, that the effect of this approach might well be that it will become impossible for institutions such as banks to obtain summary judgment in the modern age in which ‘much of their business is conducted facelessly on computer networks and recorded electronically.’

[26] We do not, with respect, agree with the approach adopted in these high court judgments. It would appear, however, that the admissibility and probative value of the respective plaintiffs’ computerised records were not specifically considered in any of these cases.

[27] There are two recent judgments of the Supreme Court of Appeal in which the respective verifying affidavits were found to be adequate, namely Rees and Another v Investec Bank Ltd 2014 (4) SA 220 (SCA) and Stamford Sales & Distribution (Pty) Limited v Metraclark (Pty) Limited [2014] ZASCA 79 (29 May 2014).

[28] In neither of these two cases, however, did the deponent to the verifying affidavit rely exclusively on the internal records of the plaintiff. In the Rees judgment the deponent also found support for her knowledge in correspondence between the plaintiff and the defendant’s attorney in regard to the defendant’s delinquent accounts. She had addressed letters of demand to the attorneys and received correspondence in response which canvassed the defendant’s defences. See para [14] of the report.

[29] In the Stamford judgment the deponent to the verifying affidavit stated, inter alia, that ‘the Applicant’s file … … which contains, inter alia, a cession of book debts in favour of the Applicant, proof of the Applicant’s claim against Quali Cool CC and all correspondence entered into by the Applicant and/or its attorney with the Respondent, is currently in my possession and under my control and I am fully conversant with the content thereof.’ See para [8] of the report.

[30] It is our view, therefore, that neither the Rees nor the Stamford judgment can be regarded as authority for or against our conclusion herein that the deponent to respondent’s verifying affidavit can rely exclusively on knowledge of its computerised records as basis for that affidavit.

First appellant’s failure to deal with the merits of respondent’s claim

[31] We should make it clear at this juncture that we do not attach any weight to the fact that first appellant did not respond at all to the merits of respondent’s claim. Counsel for respondent submitted that his failure to do so amounted to an implied admission of the claim and that this is a relevant factor to be taken into account in judging the adequacy of the verifying affidavit. He relied in this regard on the judgment of Davis J in Firstrand Bank Ltd v Huganel Trust 2012 (3) SA 167 (WCC) in which is to the effect that the contents of the defendant’s answering affidavit can be taken into account in considering the admissibility and adequacy of the verifying affidavit. Davis J said, inter alia, the following, at 177 D – E/F:

‘On an evaluation of both the claim and the defence, it could be concluded with justification that the deponent had sufficient knowledge to depose to the affidavit, which formed the basis of the factual matrix to sustain an application for summary judgment.

By contrast, there will be cases where, given the defence raised, some further knowledge is required beyond an examination of the documentation. In other words, knowledge of a personal nature may be required if it is relevant to the contractual relationship as alleged by the defendant and, if the defendant's version is proved, could constitute an adequate defence to the claim.’

[32] We differ in the first place with respondent’s contention that the absence of any response to respondent’s claim amounted to an admission thereof by first appellant. It is in our view at least equally probable that first appellant was so confident of his prospects of success in challenging the validity of the verifying affidavit that he decided that it was not necessary to respond to the merits of respondent’s claim.

[33] We do not agree, in any event, with Davis J’s approach with respect to the contents of an answering affidavit. We find the contrary view expressed by Wallis J in Shackleton para [25] logical and persuasive. In dealing with a statement, similar to that of Davis J, by Blieden J in Standard Bank of South Africa Ltd v Roestof 2004 (2) SA 492 (W), Wallis J said the following:

‘[25] Insofar as the learned judge suggested that a defective application can be cured because the defendant or defendants have dealt in detail with their defence to the claim set out in the summons, that is not in my view correct. That amounts to saying that defects will be overlooked if the defendant deals with the merits of the defence. It requires a defendant who wishes to contend that the application is defective to confine themselves to raising that point, with the concomitant risk that if the technical point is rejected, they have not dealt with the merits. It will be a bold defendant that limits an opposing affidavit in summary judgment proceedings to technical matters when they believe that they have a good defence on the merits. The fact that they set out that defence does not cure the defects in the application, and to permit an absence of prejudice to the defendant to provide grounds for overlooking defects in the application itself seems to me unsound in principle. The proper starting point is th

e application. If it is defective, then cadit quaestio. Its defects do not disappear because the respondent deals with the merits of the claim set out in the summons.’

The general nature of computer evidence

[34] By way of background it is useful to have regard to the general nature of the computer evidence that would have been available to the deponent Mr Pillay. He stated that he had personal knowledge of the relevant data regarding first appellant’s account. That information would have been available to him in in the form of respondent’s computer records, either on a screen or a printout thereof. A useful summary of three categories of computer evidence appears in a note by Stephen Mason in Criminal Law & Justice Weekly 28 September 2013 headed ‘Electronic Evidence, The Presumption of Reliability and Hearsay – a Proposal’:

‘The categories of evidence in digital format can be reduced to the following:

The records of activities that contain content written by one or more people (e-mail messages; word processing files; instant messages). It may be necessary to demonstrate that the content of the document is a reliable record of the human statement that can be trusted.

Records generated by a computer that have not had any input from a human (data logs; connections made by telephones; ATM transactions). It may be necessary to demonstrate that the computer program that generated the record was functioning consistently at the material time.

Records comprising a mix of human input and calculations generated and stored by software written by a human being (financial spreadsheet that contains human statements (input to the spreadsheet program); computer processing (mathematical calculations performed by the spreadsheet program)). It might be necessary to determine whether the person inputting the data or the writer of the software created the content of the record, and how much of the content was created by the writer of the software and how much by the person inputting the data.’

[35] For purposes of this judgment we propose to simplify the computer evidence that would have featured in the present case into two classes mentioned in the third category by Stephen Mason (see para [32] above). Each of two classes of evidence would have come into being during a separate stage of the functioning of respondent’s computer system. The first stage would have consisted of the human input, the second stage of the generation and storing of the information by the computer. The purpose of the entire process would have been to enable visual observation of the financial standing of first appellant with respondent in a complete, comprehensible and accurate form.

[36] Difficulties in regard to proof of the veracity and accuracy of the computer generated information would have arisen during each of the two stages. The human input during the first stage would have comprised the performance of certain manual operations by one or more of respondent’s individual employees and but not by Mr Pillay. The evidential problems that would have arisen during the second stage of the operation concern the reliability of the computer hardware and software that were used in the process.

The second stage – the records generated by the computer

[37] It is convenient to deal first with the evidential problems arising during the second stage, ie the generation of the records by the computer. It seems to us that these problems can be overcome by the application of a presumption of reliability. This presumption is not generally applied in the South African case law under that name but the underlying principles, we suggest, are indeed established.

[38] The presumption of reliability has a common law origin. Stephen Mason Electronic Evidence 3rd edition chapter 5 et seq discusses it under the name of the ‘Mechanical instruments: the presumption of being in order’. The Law Commision of the United Kingdom, in a report entitled Evidence in Criminal Proceedings: Hearsay and Related Topics 13.3, formulated the presumption in similar terms:

‘In the absence of evidence to the contrary, the courts will presume that mechanical instruments were in order at the material time.’

[39] In explaining the nature of this presumption Stephen Mason op cit quotes, in para [5.01], the following passage from an Australian case, Barker v Fauser (1962) SASR 176 at 178:

‘It is rather a matter of the application of the ordinary principles of circumstantial evidence. In my opinion such instruments can merely provide prima-facie evidence in the sense indicated by May v O’Sullivan [(1955) 92 CL 654]. They do not transfer any onus of proof to one who disputes them, though they may, and often do, create a case to answer. Circumstantial evidence is something which is largely based upon our ordinary experience of life. … It is merely an application of this principle to our ordinary experience in life which tells us of the general probability of the substantial correctness of watches, weigh bridges and other such instruments. If they are instruments or machines of a type which we know to be in common use our experience tells us that this is suggestive of their substantial correctness. Experience also tells us that they are rarely completely accurate, but usually so substantially accurate that people to on using them, and that subject to a certain amount of allowance for some measure of incorrectness, they act upon them. In fact, this means that for a small overweight one would necessarily …’

[40] The presumption of reliability is also applied, under that name, in the United States of America. In an article, headed ‘Old “documents”, “videotapes” and new “data messages” – a functional approach to the law of evidence’, written by DS de Villiers in 2010 Tydskrif vir die Suid-Afrikaanse Reg (Journal of South African Law) 558 and 720 he said the following, at 728, with reference to and quoting from Storm ‘Admitting computer generated records: a presumption of reliability’ 1984-1985 John Marshall L Review at 218:

‘The integrity of electronic evidence depends upon the scientific reliability of the new generation computers. The different controls, checks and tests in modern computers and software provide greater accuracy and reliability. Electronic evidence is not inadmissible simply because there is a possibility that it might be incorrect. Admissibility depends upon probabilities based on circumstantial evidence, which demonstrates that the evidence is to a certain extent reliable, not that it cannot be refuted. Although no system can ever be totally error-free, the use of security measures and certain controls can minimise these errors. Storm promotes a presumption of reliability under certain circumstances:

“It should be recognized that computer generated records carry a strong presumption of reliability which only an equally strong showing of a lack of trustworthiness should overcome. The application of the rules of evidence to computer records should incorporate such a presumption.”’

[41] In South Africa the presumption of reliability has, as far as we have been able to establish, not been applied under that name. The principles underlying it are, however, firmly established. In S v Mthimkulu 1975 (4) SA 759 (A) Corbett JA held that expert evidence as to accuracy of an instrument of measurement may in certain circumstances be obviated by the doctrine of judicial notice. See the passages at 763G-765B:

‘Whenever the facta probanda include concepts such as weight, speed, time, length (or distance), or a combination of two or more of these concepts, proof thereof must normally be presented in terms of the measures in current use at the time. This is so because it is only in terms of such measures that the concept can be communicated with any degree of precision and, in some instances, also because the factum probandum itself may be expressed in terms of such a measure. Apart from cases where human estimate is acceptable, proof of these matters of weight, speed, etc., in terms of their recognised measures, necessarily entails evidence of a measurement thereof by means of a mechanical or scientific instrument and, in some instances, in addition a process of computation.

Theoretically, such evidence of measurement should always comprehend proper testimony as to the trustworthiness of the method or process followed in order to make the measurement and as to the accuracy of any instrument used in that process.

In practice, however, the law does not always demand a strict adherence to these methods of proof. Expert evidence as to the trustworthiness of the process may be obviated by the doctrine of judicial notice. This is referred to by Wigmore in the passage quoted above. In further elaboration of the point the learned author [Wigmore, Evidence, 3rd ed, vol III] remarks (at p. 190):

"It may be premised that though, on the principle above noted, any such process or instrument must be preliminarily found to be a trustworthy one, yet, if the appropriate science or art has advanced to a certain degree of general recognition, this trustworthiness may be judicially noticed as too notorious to need evidence."

Thus, to take again the example of the X-ray, while the court may in some cases require expert testimony to interpret an X-ray photograph tendered in evidence, it does not ordinarily need to have the process of X-ray photography tendered in evidence, it does not ordinarily need to have the process of X-ray photography proved or explained to it. Judicial notice is similarly taken of other scientific instruments or processes, such as tape-recording, telephony and ordinary photography. When it comes to the reliability or correctness of the particular instrument used there is again a measure of flexibility. In certain instances the courts do not demand proof of such reliability either because of the high degree of likelihood that the machine is accurate or because it has been tested. To some extent, in this field, hearsay knowledge is, therefore, admitted and acted upon. This point is also touched on by Wigmore, op. cit., vol. II, p. 783, under the general heading of "Hearsay knowledge exceptionally admitted", as follows:

"The use of scientific instruments, apparatus, formulas, and calculating-tables, involves to some extent a dependence on the statements of other persons, even of anonymous observers. Yet it is not feasible for the professional man to test every instrument himself; furthermore he finds that practically the standard methods are sufficiently to be trusted."

Thus, to continue with the X-ray example, the court not infrequently receives evidence of X-ray photographs, in both civil and criminal cases, and invariably it does so without any enquiry being directed at the reliability of the X-ray instrument used to take the photographs.

The extent to which the court will insist upon, or relax, the standards of proof which theoretically apply when evidence involving the use of scientific instruments is presented to it will very much depend upon (a) the nature of the process and instrument involved in the particular case, (b) the extent, if any, to which the evidence is challenged and (c) the nature of the enquiry and the facta probanda in the case. No hard and fast rule can, or should, be laid down. Much will depend upon the facts and circumstances of each individual matter.’

[42] A similar approach was followed in a matter which dealt with the functioning of computers, Ex parte Rosch [1998] 1 All SA 319 (W). See pages 328h-329d of the report:

‘Chronologically the world is approaching the 21st century. Many gadgets have been invented which are capable of automatically recording material facts without human agency. Courts in this country as well as England have recognised that evidence produced by such gadgets is prima facie accurate. This accords with reality and common experience (see Wigmore on Evidence 3ed Vol 3 at 189–190, as quoted in S v Easter 1995 (2) SACR 350 (W) at 354C–355G). Some examples of cases where such evidence was found to be admissible are now listed.

In S v Fuhri 1994 (2) SACR 829 (A) the court admitted photographs produced by a machine and the information contained in such photographs, which included the digital time report. In S v Dickenson 1982 (3) SA 84 (A) the evidence produced by a gas chromotograph was held to be admissible. In R v Farden and White 1982 (1) CLR 588 (CA)and R v Dodson 1984 (1) CLR 489 (CA) the English Court of Appeal held that evidence contained in video films produced automatically was admissible in court.

As was made clear in all the cases to which reference has been made, the present case and all cases similar to it which deal with the admissibility of evidence obtained by automatic machines, relates only to the admissibility of the evidence concerned, not to the weight of such evidence. At best for the party who relies on such evidence it is open to the other party, the appellant in the present case, to lead whatever evidence he wishes in order to rebut such evidence. In our view a court would be failing in its duty if it ignored the realities of modern science and technology in the production of evidence.

In 1997 courts are entitled to accept that computers are ubiquitous in the society in which we live. The process by which these instruments record and print information is no less commonplace than the operation of motor vehicles and cameras. It is not necessary that there should be evidence as to how each computer works as a prerequisite to the admissibility of the evidence produced by such computer, if what has been produced has been done so automatically.’

[43] In the final paragraph of the passages from the Mthimkulu judgment quoted in para [41] above, Corbett JA listed the factors which may influence a court to relax the strict standards of proof. Having regard to these factors we are of the view that there are four main considerations which support the application of the presumption of reliability to the evidential problems arising during the second stage of the process, ie the generation of respondent’s computer records.

(a) The first is that respondent is a large commercial bank with branches all over the country. It can safely be assumed that its computer system is as sophisticated, efficient and reliable as those of financial institutions competing with it.

(b) It can also be assumed that respondent would employ the personnel (or outside contractors) with the experience, expertise and responsibility which the proper operation of such a computer system would require.

(c) A third factor is the relatively minor effect of a verifying affidavit in contested legal proceedings. It does not create any onus or evidential burden and it plays virtually no role in the enquiry as to whether a defendant raises a valid defence in its answering affidavit.

(d) A fourth factor is that respondent’s computer records with respect to any account are accessible to the client. Statements are sent to the client and information may, for example, be accessed by telephone, through ATMs (automated teller machines) or via internet banking. This aspect would tend to minimise the effect of possible mistakes.

The first stage of the process - human input

[44] The evidential problem that arises during the first stage of the process is that the human input is hearsay evidence. The probative value of the operations performed by the persons in question depends upon their credibility as witnesses. There are, however, two ways, or a combination of both, by which the hearsay problem may be overcome. The first is that it may be admissible as an exception to the hearsay rule in terms of s 3(1)(c), read with s 3(4), of the Law of Evidence Amendment Act 45 of 1988 (‘the statutory hearsay exception’). These provisions read as follows:

‘3 Hearsay evidence

(1) Subject to the provisions of any other law, hearsay evidence shall not be admitted as evidence at criminal or civil proceedings, unless-

… …

(c) the court, having regard to-

(i) the nature of the proceedings;

(ii) the nature of the evidence;

(iii) the purpose for which the evidence is tendered;

(iv) the probative value of the evidence;

(v) the reason why the evidence is not given by the person upon whose credibility the probative value of such evidence depends;

(vi) any prejudice to a party which the admission of such evidence might entail; and

(vii) any other factor which should in the opinion of the court be taken into account, is of the opinion that such evidence should be admitted in the interests of justice.

… … …

(4) For the purposes of this section –

“hearsay evidence” means evidence, whether oral or in writing, the probative value of which depends upon the credibility of any person other than the person giving such evidence.’

[45] There is support for such an approach in the case law. Although the judgments mentioned hereunder dealt with the concept of a data message as defined in ECTA the reasoning therein would apply with equal force to the computer generated documents that we are dealing with in this case. In Ndlovu v Minister of Correctional Services and Another [2006] 4 All SA 165 (W) at 174e-175c the court was presented with computer printouts reflecting the monitoring of the plaintiff by the prison authorities. The judge (Gautschi AJ) held that the information that was recorded by persons who did not give oral evidence, was hearsay evidence. He admitted it, however, in terms of the statutory hearsay exception.

[46] In S v Ndiki and Others 2008 (2) SACR 252 (CkHC) Van Zyl J dealt, inter alia, with the admissibility of computer generated documents. In obiter dicta at paras [31] to [33] of the judgment, he expressed the view that the documents constituted hearsay evidence which could be admissible in terms of the statutory hearsay exception.

[47] In LA Consortium & Vending CC t/a LA Enterprises v MTN Service Provider (Pty) Ltd 2011 (4) SA 577 (GSJ) Malan J, writing on behalf of a full court, held that a data message may contain hearsay which would only be admissible insofar as it passes the criteria set out in the statutory hearsay exception.

[48] Having regard to the various factors set forth in the sub-paragraphs of the statutory hearsay exception we are of the view that the hearsay evidence regarding the human input in the present case, should in the interest of justice be admitted as evidence. Apart from the considerations mentioned in para [41] above, it would be practically impossible to obtain affidavits from all the persons that participated in providing the human input. Many of them might not even be identifiable when the verifying affidavit is deposed to. The admission of this hearsay evidence would only prejudice first appellant if he has no defence to respondent’s claim.

[49] An alternative solution to the problem regarding the admissibility of evidence arising from the human input is the application of a presumption of regularity. This presumption has been described as follows in Zeffert v Paizes second edition The South African Law of Evidence at 212.

‘The scope of the presumption of regularity, usually expressed in the maxim omnia praesumuntur rite esse acta, is very ill-defined… … In some cases it appears to be no more than an ordinary inference, based upon the assumption that what regularly happens is likely to have happened again. In other cases it is treated as a presumption of law, sometimes placing an onus upon the opposing party and sometimes creating only a duty to adduce contrary evidence. It has been applied in a wide variety of cases which are impossible to catalogue exhaustively.’

[50] The presumption of regularity was applied by Steyn J in the context of the operation of computers in R v Minors; R v Harper supra at 213:

‘Moreover, in a great many cases the necessary evidence could be supplied by circumstantial evidence of the usual habit or routine regarding the use of the computer. Sometimes this is referred to as the presumption of regularity. We prefer to describe it as a commonsense inference, which may be drawn where appropriate.’

[51] There is clearly a significant degree of overlap between this presumption and the presumption of reliability discussed above. This is understandable as the reliable operation of a computer also depends upon the quality of the human input. It seems to us therefore that the arguments in favour of the application of the presumption of reliability in this case, mentioned in para [32] above, applies mutatis mutandis to the application of the presumption of regularity.

Returning to Maharaj

[52] We revert finally to the judgment of Corbett JA in the Maharaj case. We believe that our approach herein is not inconsistent with the principles applied in that judgment. Corbett JA accepted, for pragmatic reasons, that the manager of the branch of the respondent bank who deposed to the verifying affidavit could not have been expected to have personal knowledge of every entry in the client’s statement of account. He ‘must needs rely upon the bank records which show the amounts paid into his account and the amounts withdrawn by the client’. The learned judge of appeal did not, so it would appear, provide any express indication of the legal basis for this approach. It seems to us, however, that it was not inconsistent with the application of the presumption of regularity.

[53] The technological environment was in any event very different from what it is today. The Maharaj judgment was delivered in 1975, before the advent of the information revolution referred to above. Had computers, as we know them today, been used in the ordinary course of banking business, Corbett JA might well have applied the approach which he himself articulated in the Mthikulu judgment. The latter judgment, incidentally, was delivered by him only two months before that in the Maharaj case. It is also relevant that the statutory hearsay exception did not exist in 1975. For that reason the recognition of the human input as hearsay evidence would not have solved the problem regarding its admissibility.

[54] It may also be noted in this regard that the terms ‘personal’ and ‘direct’ which appear in the passage at 423BC in the Maharaj judgment, quoted in para [20] above, do not appear in Rule 32(2). In terms of that rule the deponent is only required to ‘swear positively’ to the facts in question. Mr Pillay would have been authorised to have access to respondent’s computer records and he would have been qualified to understand and interpret them. He would therefore have been in a position to depose to a verifying affidavit that complied with Rule 32(2).

Conclusion

[55] We are accordingly of the view that the computer generated information of first appellant’s financial standing with respondent that was available to Mr Pillay, was sufficient to allow him to depose to a valid and adequate verifying affidavit. We find that he in fact deposed to such an affidavit.

[56] First appellant’s appeal thus falls to be dismissed with costs. On 27 June 2014 respondent abandoned its judgment against second and third appellants. Respondent is, however, responsible for the costs of the appeal incurred by second and third appellants up to 27 June 2014.

[57] In the result, we make the following orders:

(1) First appellant’s appeal is dismissed with costs.

(2) The granting of summary judgment against first appellant by the court a quo is confirmed.

(3) Second and third appellants are given leave to defend the action.

(4) Respondent shall bear the costs of the appeal incurred by second and third appellants up to 27 June 2014.

(5) The costs in respect of the summary judgment application against second and third appellants shall stand over for later determination.

BLIGNAULT J

Judge of the High Court

HENNEY J

Judge of the High Court

GOLIATH J: I agree.

GOLIATH, J

Judge of the High Court

---