RACV Insurance Pty Ltd v Unisys Australia Ltd [2001] VSC 300

RACV Insurance Pty Ltd v Unisys Australia Ltd [2001] VSC 300 - VSC 2001 case summary — Zoe

[10]

It is appropriate to make some reference to the nature and size of the trial which took 32 sitting days. The trial was conducted on witness statements with a court book. By the end of the trial the plaintiffs had called 10 witnesses and Unisys 12 one of whom had not provided a witness statement. In addition, six witness statements and one affidavit were received into evidence by consent without the necessity to call the witness. There were three witness statements from one witness and two from another. The evidence was of a factual and technical nature concerning the computer system. To the close of evidence the transcript ran to 2,731 pages. It extended to 3,178 pages by the close of final addresses in which counsel spoke to written submissions which ran to 378 pages and five supporting files of authorities. Then, six weeks later, on 15 May 2001 Unisys applied for leave to amend its defence to raise a new point; following a hearing the application was withdrawn (see [54] below). At the outset of the trial the court book consisted of 49 lever arch files (vol. 1 devoted to the pleadings) which contained 18,935 pages. I also was given another lever arch file which contained the December 1993 agreement and which seemed to comprise several hundred pages at least. When finally tendered at the conclusion of evidence the court book was reduced, by the agreed discarding of thousands of pages, to 28 volumes including vol. 1 containing the pleadings. I am, however, left with many thousands of pages of oral and documentary evidence. In the practical world, in which other litigants are prevented, by litigation such as the present indulged in by parties with deep pockets, from getting their cases on, or whose cases are refixed because of limited judicial resources, there must be a limit on the extent of the time I can take to discuss and analyse the documents and the evidence in this judgment, let alone scour the thousands of pages to see what is there to which little or no reference has been made. I am, as I must be, guided by counsel's final addresses in identifying the issues and the relevant materials. I emphasise though, that even then, they did not at that stage refer to every possible piece of evidence. That is for the obvious reason that it was impractical to do so and because so much of it was not necessary to refer to for the resolution of the issues as they were finally left for determination. I do emphasise that in arriving at these reasons for judgment I have read and considered, subsequent to their making, all that counsel said in final addresses, both oral and written, and in their opening addresses, and the evidence of the witnesses.

[25]

(a) Robert Alexander Mitchelson who was general manager of RACVI at all relevant times until he retired in December 1995. He was the senior management person responsible to the board of RACVI and to the chief executive officer of RACV. He signed the Unisys contract on behalf of RACVI.

(b) David Charles Hurford who from 1992 was assistant general manager of RACVI reporting to Mitchelson and in January 1996 was appointed general manager following Mitchelson's retirement. He continued in that position until 30 March 2000 when he retired from RACVI.

(c) Paul Arthur Lang who was employed by RACVI as claims services manager between September 1992 and December 1996 inclusive when he retired from RACVI. He reported to the general manager.

(d) Lisa Joy Chesson-Kistas who was employed by RACVI as a claims consultant in the motor vehicle section. By reason of her expertise in the claims area she was involved in the project from 1992 and on a full-time basis from at least November 1993 to its end in June 1996.

(e) Craig Donald Haw who as an employee of RACVGS worked full-time on the project from February 1994 until it ended in June 1996. He was responsible for technical input in respect of the interfaces between the Unisys workflow and electronic imaging system and the existing RACV computer system. He was not responsible for any aspect of the Unisys system including its design, functionality and performance. He is now employed by RACV as Unix administrator.

(f) Justin Taylor Doherty (formerly Clack) who was employed by RACVI from 2 February 1993 until 15 August 1997. He was project assistant and later project operations manager for RACVI during the implementation of the Unisys WMS system. He was responsible for RACVI's day to day operational issues relating to the WMS. He worked full-time on the project from August 1994 until the start of July 1995 except for November 1994 and from July 1995 to end June 1996 was almost full-time.

(g) Dennis Hakme, who as an employee of Memorex Telex conducted an analysis of the LAN in the March to May 1996 period.

(h) Peter Malcolm Jenkins who was finance manager of RACVI in 1996 and is now employed by RACV. He produced accounting records of RACVI relating to and recording the payment of various accounts and wages information to establish items in the plaintiffs' particulars of loss and damage.

(i) Gregory Errol Blashki, an independent accountant, who as an expert gave opinion evidence to establish a general increase in labour costs suffered by RACVI as a result of the Unisys contract.

(j) Peter George Thorne, a consultant and a principal fellow in the Department of Computer Science and Software Engineering at the University of Melbourne. He was formerly head of the school of computer science located in the science and engineering faculty and deputy dean of the latter, although having moved from being a full-time staff member to more of an associate capacity with some teaching and research, he retains the rank of reader and associate professor.

(k) The following three witnesses whose evidence in witness statement form was tendered by consent without the need to call the witness:

[30]

(a) Craig Mitchell Felsenthal, an employee of Unisys, who in 1993 was customer relationship manager, and who was the first at Unisys to become aware of the RACV project.

(b) Brian Patrick Josephson, who was employed by Unisys from April 1987 until May 1996. He became the marketing manager for InfoImage shortly before its launch in October 1989 and in 1994 and 1995 was the lead sales specialist for the product. His involvement in the RACV project commenced in early 1993 and concluded shortly after the Unisys contract was signed.

(c) Ruth Elizabeth Willis, an employee of Unisys with a background in mainframe development of software. She had no experience with the InfoImage product when she became involved in the RACVI project. She was bid manager for the production of Unisys' response to RACVI's request for proposal and worked on the bid until the contract and subsequently as a project manager to manage the environment (but not the technology) including risk and issue management. She scheduled staff to perform the technological aspects of the implementation, gave tasks and a completion date and checked what they had done. Her involvement ceased in April 1995 following RACVI's acceptance of the system. She is no longer employed by Unisys.

(d) Vincent Joseph Wall who from 1994 to 1996 was employed by Unisys in the management consulting group to work in imaging and workflow. He was involved in several aspects of the RACV project from the commencement of his employment in early 1994 until sign-off in March 1995.

(e) Neil John McFarlane, an employee of Unisys who was involved with the WMS project at several points. Pre-contract he provided advice as to the capacity of the LAN at RACV to support a WMS system which used InfoImage; and in mid 1994 and again in or around March 1996 he was called into to consider the LAN.

(f) John Thomas Chmura, who for some years until 5 December 2000 was an employee of Unisys and its predecessor companies. In his last position he was the systems architect specialising in image document management, workflow and records management and under various titles had worked in that area of computer processing since 1991. His involvement with the RACV project was to run three courses for RACV and Unisys staff at the Unisys office and at the request of Willis work full-time at the RACV from 5 September to 7 November 1994 as a consultant in building the InfoImage solution taking into account the customised functionality. This involved defining the InfoImage application for RACVI, that is, telling the InfoImage system how to behave in doing its work. He taught two more courses late in November 1994 to RACV and Telstra personnel, after which his involvement was almost ended. He returned to RACVI in June 1995 to review problems with the system.

(g) Geoffrey Edmund Olsen, an employee of Unisys between 1991 and 1999. He worked on the project from the first to the last. He assisted in the demonstration in May 1993 and thereafter worked on technical aspects as a member of the team and in that respect on the configuration of the system to be installed and following the award of the contract he was engaged full-time on the implementation of the InfoImage system until the project was terminated in June 1996. He gave evidence over five days.

(h) Kenneth Ross Johnson who commenced employment with Unisys on 23 May 1994 as a consultant for the imaging services division. He said he was employed with a view to becoming a systems administrator for RACVI system and as such be on site during the warranty period. He was assigned to the RACV project in June 1994 at RACVI's office, read materials on InfoImage and attended lectures by Chmura, and performed a variety of tasks in developing the system. In August 1995 he attended a meeting at Plymouth in the United States to discuss the rebuild of the system. He continued to be involved up to termination of the project in June 1996.

(i) James Joseph Esmonde, who since December 1994 has been employed by Unisys as the service delivery manager in Cross Industries. He became involved with the RACVI project in May 1995. While Birthisel was on the technical side, Esmonde worked on the business management side. His role, he said, was not "getting involved deeply into the technical aspects of products; it was to ensure that we had resources available to do that sort of thing". An instance is that he requested Birthisel be appointed to project manage the team. He also performed a liaison role with RACVI. He was involved to termination in June 1996.

(j) David Vernon Birthisel, a Manager in Unisys' systems integration practice who has been employed in various roles by Unisys since March 1986. His involvement with the RACV project commenced in May 1995 when he succeeded Willis as Project Manager. He reviewed the system that had been accepted in March 1995. He recommended that RACVI stop adding new claims (which meant handling them on the old manual paper-based process) and shortly thereafter recommended that the Unisys system be abandoned altogether and that RACVI fully return to the manual system, and that the Unisys system be rebuilt. He was involved through the subsequent rebuild of the system until the termination of the project in June 1996.

(k) The following witnesses whose evidence in witness statement or affidavit form was tendered by consent without the need to call the witness:

(i) Kevin d'Souza who has been employed by Unisys since 1993 as a technical specialist with a background in designing and developing operating systems. In a short witness statement he referred to specific involvements with the RACVI project, initially at the request of Willis in or around late 1994 when over a couple of months he completed some self-contained tasks in conjunction with Wall and Olsen, and some time later at the request of Birthisel.

(ii) Daniel Rouqueirol who has been employed by Unisys since 1986 and is a program manager. His involvement commenced in early 1994 when he was primarily responsible for getting a team together to work on the RACVI project. He chose people who had been involved in the pre-sale, and who included Willis, Olsen and Wall. His short witness statement referred to a few matters and said his involvement ended when the system was accepted and went live around April 1995.

(iii) William H. Maier, a former employee of Unisys Corporation who retired in November 2000. He had been a software support manager since 1991. He described the support group system in the United States and how it operates including sending analysts to customer installations overseas and, after a brief reference to the number and range of customers using InfoImage and to the Unisys or NT platform, he referred generally and with somewhat broad observations, to the RACV project.

(iv) Michael Aaron Bromberg who, as associate general counsel for Unisys Corporation, produced the following copy licence agreements: an agreement between Unisys Corporation and FileNet Corporation dated 3 August, and an agreement between Unisys Corporation, Sigma Imaging Systems Inc. and Empire Blue Cross and Blue Shield Inc. dated 17 March 1990.

(l) Brian Bertram Hesom, an independent consultant with experience in computer technology including process redesign and the implementation of new processes. He gave opinion evidence as to a report he had prepared (which was tendered) in which he answered a series of questions put to him by Unisys' solicitors.

(m) Duncan Nicholas MacCallum who is now a partner in Deloittes Consulting but who at the times relevant to this case was employed by Deloittes as a principal consultant and whose expertise was in information technology. In August 1992 Deloittes was engaged by RACVI to provide consultancy services in the re-engineering of the process and systems in its motor claims business. MacCallum was the lead person for Deloittes. He was involved both pre and post-contract although not continuously in the post-contract period. He gave evidence on a variety of matters. A witness statement had not been provided for him as a result, as I understood it, of Deloittes having settled on confidential terms and his reluctance, however inspired, to confer with Unisys' lawyers.

[104]

"Workflow Management Architecture The major considerations in establishing the architecture for the provision of workflow management facilities to a centralised claims process, with some supporting activity in the current branch environment, are: * Performance and reliability * Security and integrity of data, indexes, process scripts, etc. * Ease of end user operations * Communication infrastructure * Development and database environments. These factors, combined with the centralisation of claims and the need to ensure consistent work practices, suggest that a centralised workflow management architecture is the most appropriate solution." "Image Services Architecture The major considerations in establishing an image services facility are: * Retrieval and distribution requirements * Performance and reliability * Security and integrity of data, indexes, and images * Communication infrastructure * A central image storage facility with connectivity to the Amdahl mainframe host application and/or the Workflow management System * Role of optical and magnetic storage." "Workflow Management Functionality The Workflow Management environment must provide for development and execution of automated workflow paths or scripts. The following functions are considered minimal: * Application Programmer's Interface (API) and user exits (or, remote procedure calls) * Data Manager, or, standard relational database product with SQL (Structured Query Language) interface * Portability of operating software (client and server) * Activity, or, workflow map builder * Workflow simulator for performance analysis * Status monitor for displaying the status of work flowing through the system * Audit trail to track the history of activity in the system * Case and folder management functions * Queues definition * Activities definition * Routing - dynamic - guaranteed - parallel - rendezvous * Support of rule-based routing * Management reporting utilities * Support for forms and screen design * Recognition technology for machine printed characters or codes on forms * System administration for users, groups, authorities, security and other operational requirements. Image Service Functionality The Image Services environment must provide the following minimum functionality: * Image capture (scanning and committal) of varying documents, including double sided multi page * Availability of automated indexing techniques * Management of colour photographs * Support minimum scanning resolution of 200 dpi * Support background drop-out * Input/output via a fax gateway * Standard relational database manager for document indexes * Association of annotations with images and folders * Full audit trail of folder and document activity * Ability to associate text and data object with folders * On-line storage of all current, active claims, and near line storage of all inactive claims within a pre-determined time frame * Security profiles for image access * Optimisation techniques for retrieval performance * Multiple document display * Image caching at client and server * Open access from mainframe host applications and the proposed Workflow Management System." "Integration It is mandatory that the Workflow Management System and Image Services Facility are capable of integrating with RACV's existing applications environment."

[179]

I turn to the defendants' witnesses. It is convenient to commence with Josephson. He expanded on his witness statement in oral evidence in chief. He remembered attending presentations of the InfoImage product to RACV but did not specifically recall attending a meeting at this time (4 August 1993). He recalled that as a result of advice of implications of the Trade Practices Act he was cautious about what he said. He said that the demonstration to RACV showed a document retrieval occurring within a 2-4 second period. He was certain he said that the documents were being retrieved in a short time period because the image was being retrieved from magnetic disk. He said he explained that the retrieval time would depend upon the media in which the images were located. He did not deny saying that the response time demonstrated was indicative of response times that the RACV could expect for on-line claims, however he said this could only be expected for on-line retrieval (that is, retrieval from cache) and would not be expected from optical disk. He did not say that all images would be stored in cache. He did not deny saying that Unisys had technical expertise to implement the InfoImage system. He did not recall saying that Unisys could meet RACVI's requirements and would show that in the demonstration or that Unisys considered the InfoImage system was ideal for RACVI's requirements and would meet its business needs. He did not specifically recall Chesson-Kistas saying that one of the most significant issues for RACV was locating a file when dealing with claims on the telephone or that it was important to locate and access a file very quickly. He said (by way of explanation but not as recounting a discussion with Chesson-Kistas) that retrieval was only one part of the InfoImage technology which enables business process re-engineering, that it was possible to access data to deal with telephone inquires without looking at an image, and that the ability to view an image was only one component of the entire package and successful implementation of the product required a commitment to business process change by RACV.

[181]

To the first question in cross-examination he said that he believed that conversation occurred at the first demonstration. I interpolate by way of reminder that that was in May 1993. He was then asked to state the retrieval times that he gave in the conversation. His answer was: that retrieval from PC cache was sub-second, generally speaking, that is what it is capable of; 2-4 seconds from cache on the server; up to 20 or 30 seconds from optical disk off-line; and from optical disk on-line would be much quicker and depending on the activity on the system, the number of users and the final configuration, could be up to 10 seconds. Much later the cross-examination returned to this conversation and Josephson said that he probably explained what a jukebox was, that there were different sizes, and referred to response times from the jukebox. Further on, he said that he regarded retrieval times as important and that he volunteered the information without being asked. Later he said that his advice as to retrieval or response times was said at the demonstration meeting "whenever that was" and that he considered that significant to selling an InfoImage system. Later again, almost at the end of his cross-examination Josephson was taken to his witness statement and to that of Chesson-Kistas dealing with the 4 August demonstration. I accept that he recalled there was a demonstration after the July response, and another which Mitchelson attended, but he could not recall dates. His evidence was they were all important in the process of winning the contract. With the following exceptions he agreed with the words or substance of Chesson-Kistas's version. The exceptions were that he did not say that the retrieval time from near line was the same as on-line, he said near line was 20-30 seconds approximately; and, in relation to para. 60 he did not recall Chesson-Kistas saying the prime need was to access the file very quickly or that he said the system would meet the requirements for timely customer requirements. He knew they discussed those topics, "but not like that".

[209]

Mitchelson wished to be satisfied for himself that the system would perform the tasks stated and meet RACVI's business needs. Accordingly, at his request, MacCallum arranged for Unisys to demonstrate the system to senior management. The primary purpose was to demonstrate the system to Mitchelson who was to make the final decision. Mitchelson also viewed the video provided by Unisys called Document Imaging which indicated improvements to the claims system and saving of costs. The demonstration occurred on 2 December. Lang attended the presentation; he said it was in effect an enhancement and repeat of the previous demonstrations. Josephson and Willis were present and gave some evidence about it; others from Unisys were also present and involved in the demonstration. The agenda shows that the meeting was to occupy 1 ½ hours. According to Mitchelson, Josephson, who addressed the approach of Unisys to RACV's requirement, said that Unisys proposed to replace the manual handling of documents by electronic images and electronic claim forms using the InfoImage system which was suitable and ideal for RACVI's purposes. He spoke about the partnership between RACVI and Unisys and said that Unisys had professional project implementation experience second to none. He referred to the implementation at Telecom. He talked about the InfoImage product, saying the architecture was ideal, flexible, fitted in well with existing environments and was proven. Kumar gave a prototype demonstration which illustrated the life cycle of a claim. It showed how documents were converted to images. Documents were moved from screen to screen, and retrieved from memory, simultaneously. Documents were created simulating a telephone claim. The system was easy to use and stable. Mitchelson was very impressed; the system looked what was needed to streamline RACVI's claims processes and meet its business requirements. Walker made a presentation on how the project would be managed. The teams and structure of authority were shown on the slides. Mitchelson was impressed that a professional team had been put together. I do not accept the evidence of Willis that this was not put forward as a commitment to make the nominated people available. Mitchelson said that Unisys personnel expressed no qualification about its ability to produce the system required by RACVI, including as to response times. Willis gave evidence that she did not recall any statement regarding response times. She said that Unisys sales personnel are instructed not to make such statements; that evidence was not correct and I reject it. The truth is, as she stated in cross-examination, that sales personnel are instructed not to make statements they cannot deliver on. That is an entirely different thing. They were not instructed never to make a statement regarding response times. I add further that I accept Mitchelson's evidence; he impressed me as an honest and reliable witness. Overall he was impressed with Unisys and the product. He agreed with the recommendation of the Team and that the matter should move forward with Unisys.

[220]

Clause 2 is a definition clause. "Product" was defined to mean the equipment, software and related documentation. "Equipment" meant the hardware itemised in the schedule; "Software" meant the objective code version of computer programs identified in the schedule and any related documentation, and includes customised software; "Customised Software" meant software or modifications written in accordance with an agreed set of functional specifications; "Functional Specifications" meant the specifications agreed for the customised software which was referred to in the schedule or written by "us" pursuant to cl. 5(a); and "Related Documentation" meant all materials other than the software supplied by Unisys including operating instructions. The word "System" meant the software and equipment. "Services" meant all forms of installation, integration, maintenance services for the system and training as noted in the schedule. "Maintenance Services" meant the equipment maintenance and software support services identified in the schedule. "Requirements Definition" meant a document written by RACVI or jointly with Unisys describing the functionality required for the customised software. "Acceptance" meant the system has passed the acceptance tests, and "Acceptance Test(s)" meant the mutually agreed tests based upon objective criteria designed to test the compliance of the products and/or system with the functional specifications and/or published specifications and which will be referenced in the PMP. It is to be noted that in section 2.4 of the PMP, which is concerned with acceptance, it is stated that the test plan for acceptance testing will be jointly prepared by RACVI and Unisys, and that acceptance testing to prove the capabilities of the system in accordance with the contract must be performed in a planned and controlled manner in order to quickly and effectively confirm the suitability of the system for production. It was also stated (in section 2.4.2) that RACVI will develop acceptance test procedures and a plan that proves that acceptance criteria are satisfied with assistance from Unisys.

[253]

From 28 August to 1 September 1995 a high level meeting was held in the United States with personnel from Unisys Corporation and Filenet with Sigma and Wang personnel contributing by telephone. Johnson, d'Souza and Del Gannon (project director on the RACVI account) attended from Unisys. Johnson gave evidence about the meeting. In his witness statement he said the meeting was to discuss the rebuild of the WMS at RACV. A document was prepared as an "informal recap" of all the issues discussed. It stated that the objective of the taskforce was to: define the problem, and differences between the standard product and RACV implementations; identify short and long term alternative to stabilise the system and make it supportable; review existing and proposed design for stability and long term supportability; suggest design alternatives, given constraints imposed by the project schedule, particularly for long term supportability; validate the design with Wang/Filenet and ISD personnel; identify the risks and means to manage the risk. Under Findings and Recommendations it was noted that the team worked under the basic constraint that any short term solution must be implemented by November 1995. The existing solution had several major shortcomings, most of which were mitigated by a new design presented by the RACV team. The new design was more efficient from a performance perspective. Johnson's evidence was that the part of the InfoImage product called event manager did not provide the essential functionality needed to deal with situations created when users did not follow the defined processes, and the object of the visit was to get permission from event manager experts to build their own unique processes. The source code to event manager was obtained which then permitted the functions to be built using existing functionality. The notes identified the major risks in the plan as: schedule risk as the integration and stress testing time (four weeks) leave little room from recovering from any major errors; hardware under configuration risk, as there is no easy way to model the new design; related to that, the adequacy from a performance perspective is unknown. I interpolate that in cross-examination Johnson said that the reference to performance was not to retrieval times as the focus was primarily on how to make event manager more efficient and in that respect he referred to moving folders out of the diary queue to the claims officer in a more timely manner. Then, in a section that listed root causes of existing instability and performance issues, it was stated that no capacity planning or performance analysis has been done on the existing system, so performance is a big risk when the customer goes full volume. I interpolate that in his evidence Johnson said that at that time there was no way to measure how the system would react under load. Further on in the notes, under the heading Performance, it was stated that the existing solution is manifesting itself as a major performance issue for the customer, the main performance perceptions for the customer being: diary documents not sent to the desktop on the diary date, but sometimes they are a day late or not sent out at all; it may take more than one minute to file a document into a folder; it can take several minutes to retrieve a folder onto a desktop. In his oral evidence Johnson said that that reference was related to getting things out of the diary, getting folders back to the user, into their queue, so they could retrieve them, whether the user was on the telephone or had received correspondence. That was "causing a lot of problems". In his witness statement Johnson said that these matters were discussed and that the performance issues were addressed during the rebuild. Johnson gave some elaboration of these matters in cross-examination.

[299]

The matter was discussed at a meeting on 12 April which Hurford and Walsh attended. Walsh said the project was on target. The matter of system performance criteria was discussed and Doherty's memorandum of 27 March and the Unisys response dated 11 April were tabled. I accept Hurford's evidence that he told Walsh that the response times for retrievals set out in the RFP were mandatory and that other response times could not be accepted. In evidence he said that he thought that what was set out in the 11 April 1996 memorandum reflected the RFP. And that while he used the word "mandatory" a further second or so would not have caused him difficulty. Hurford said that neither at the meeting nor at any time was there a discussion he could recall in which any senior Unisys person said there were no performance criteria in the contract. The whole basis of the business case for RACVI going into the system was that it would achieve the sort of outcomes set out in the RFP. Hurford said that the thrust of the conversation was about the types of retrieval times viewed by RACVI as being necessary for the system to pass the test. A further meeting was held on 22 April. Walsh said the project was on track. Hurford told Walsh that he required the system to perform in accordance with criteria in the original documents and nothing less was acceptable. There was to be further testing. There are notes of the meetings. They reflect that Unisys was to respond to the performance requirements of RACVI, then the effect of performance on business functions was to be discussed; then it was identified the system performance could not meet four criteria; then there was to be tabled a document stating required performance levels, a business case and estimated possible performance, to be updated in light of stress testing and alternative approaches were the system unable to provide sufficient performance. As to these notes of meetings, which were not signed as verification of accuracy or acceptance, I find that they indicate discussion but I do not accept that RACVI or Hurford in particular went along with a proposition that RACVI would accept the Unisys system even if at the time of delivery it could not meet "the four criteria that current system performance cannot satisfy" and that long term options be considered after production rollout. It seems that it was hoped that by 29 April(when there was to be another meeting) required performance levels for individual transactions would be established.

[331]

Volume/stress testing was carried out on Saturday 1 June. The test was conducted between 10:00 a.m. and 12:36 p.m. with 31 participants, the object being to simulate an environment of maximum load by means of multiple users moving multiple documents around the system. That was the most individuals who had been on the system at one time. The test was supervised by Doherty. At 12:36 p.m., while the test was being conducted, a `system failed' message appeared on all PC's. Testing was halted for one hour to allow Unisys to discover and resolve the error. After discussion with Birthisel the test was stopped. Unisys took a dump of the system at the time of the error and sent it to engineers in Sydney and the United States. The evidence of Birthisel and Olsen is that it was a Unix problem. On 7 June Unisys installed a patch to resolve the problem. During the test response times were taken over a number of transaction types. In Doherty's view they indicated that performance of the system remained unacceptable. There is a record of response times at pages 9930-9933 in the court book. In a memorandum to Hurford dated 6 June Doherty reported that the response times for the retrieval of folders remained in the range of 20-40 seconds on average, with the occasional retrieval taking up to two minutes. He noted then that the response times for other transactions are still to be collated. That collation may be in the above record of response times. He noted that claims staff continually commented on the slowness of the response times. Chesson-Kistas, who was present during the test, said that the response times got slower as the test progressed and that they were unacceptably slow. In his report Doherty also referred to a problem on functionality of the creation of multiple copies of a document in a folder. The PC had to be switched off to stop it happening. That had occurred when the system went live in 1995 but had not been observed through the extensive testing in the past two months. Unisys had referred this to Sigma in the United States. An error or exception report was raised for that and another matter. I note the evidence of Chesson-Kistas that on 26 June 1996 Birthisel told her that Unisys was awaiting a response from the United States about the problem of the system producing multiple copies of documents.

[332]

On 5 June Hurford met with Walsh and Felsenthal. The meeting was to see if Walsh and Hurford could find common ground further to their letters dated 23 and 31 May respectively. Hurford made a note of matters discussed. He spoke about the expectations of RACVI, that the WMS system produce tangible benefits of increased productivity, staff reductions and reduced costs of claims processing, and a system in which claims files were found more quickly than if an operator had to go and find them in a filing cabinet. There was no advantage in a new system that was not quicker than manual retrieval. Walsh said that Unisys maintained its high level of confidence the system would meet RACVI's business requirements. He said that might not happen on 24 June but certainly at a later date and with some enhancements, the cost of which would need to be discussed. In cross-examination Hurford said, and I accept, that the reference to enhancements was not elaborated by Unisys and he was not sure what they were coming at. The solution to the 1 June problem was awaiting analysis in the United States (as stated above, a patch was effected on 7 June.) Walsh also acknowledged there were errors of importance awaiting solution, including the general protection fault. Hurford's notes go on to record that the parties agreed to work towards a definition of "success" and would together answer the question: what can be achieved by 24 June for the implementation of a production model? Hurford said he was looking for a situation in which he was able to go into production, perhaps being less rigid, but "success" was never defined. There was reference to the trial of a more powerful PC to determine if the performance criteria could be improved, Hurford pointing out that RACVI had installed equipment advised by Unisys. Hurford said that the situation had to be resolved before 24 June, and that he had to report on the outcome to the board on 26 June and this would be to abandon the system altogether and pursue compensation, or use the system wholly or in part. Walsh agreed to state in writing what could be achieved by 24 June, in particular if timely retrieval could be made as required.

[359]

"Dear Justin, Further to our discussion yesterday regarding outstanding system performance issues, I would like to outline a go-forward plan that will allow us to work towards a satisfactory resolution. The four performance standards, where major variations between RACV requirements and existing system capability limitations have occurred, have been categorised as Filenote Retrievals and Folder Retrievals. With the research we have conducted to date, including input from sites in Austria and the US, we do not believe there are opportunities for major improvements to Folder retrievals, due to hardware and software constraints. I understand David Birthisel has discussed the use of pre-fetch processing with you to assist in providing consistent performance regardless of the number of expected users. We will therefore investigate further the use of pre-fetch as a means of sustaining a 20 to 40 second retrieval time. There has been preliminary discussion on various options to improve Filenote retrievals. Before Unisys is in a position to recommend the most appropriate option, we propose to conduct a technical feasibility study, commencing on Wednesday of next week. The study will be undertaken by Geoff Olsen and Leon Pearl and we would like also to obtain input from your IS team to assist with the assessment and review process. It is expected the work will be completed within two weeks from the start date, given current commitments. We anticipate being able to significantly reduce the current performance in this area and expect to provide more specific information at the conclusion of this study. The objectives of the study will be to assess various options, to determine the likely impact on existing functionality and performance and to estimate the effort required to specify, develop, test and implement the solution. Unisys is prepared to conduct this exercise at nil cost. However, before progressing to the development stage, we will need to confirm funding arrangements. I would appreciate confirmation of your acceptance of the above and the availability of an IS resource to participate in the feasibility study."

[364]

A volume/stress test was conducted on Saturday 15 June between 9:00 a.m. and 3:00 p.m. with 23 claims officers participating. It was supervised by Doherty whose report is in the Court book at page 9935. Three personnel used stop watches to record retrieval times. Doherty included these times in his report which he said, and I accept, correctly recorded events on that day. The report refers to 157,808 documents as "total documents processed". The word "processed" may be misleading: it refers to the number of documents in the Oracle database of the system and not to the number of documents retrieved to the desktop of PC's during the testing (see the witness statement of Ben Connell, para. 4). The report stated that no Unisys hardware/network related problems were observed. Three functionality exceptions were raised (one concerning the problem of the creation of duplicate documents) and analysis of queue contents after the test disclosed two matters, all of which issues needed to be addressed prior to acceptance of the system. Tables of response time recordings show that in the case of two transaction types, retrieval of folder from filing cabinet and retrieval of folder from inbox, timings were taken prior to 10:00 a.m., between 10:35 a.m. and 11:40 a.m., and between 12:45 p.m. and 13:45 p.m. A deal of evidence was given as to the reason for variations in the timings thus obtained. The recorded response times for those transaction types, and the transactions retrieval of scanned documents from inbox and open batch on desktop, were appreciably slower than the expected response time: the other five transaction types matched or were in the order of the expected time. In relation to the retrieval of folders, the report noted that the retrieval times for folders from the inbox were significantly higher prior to 10:00am and that at approximately 10:00am adjustments were made to the positioning of optical platters which resulted in a significant decrease in response times. The report noted that a more detailed report of the need for this work is required from Unisys. It also noted that the response times for folder retrieval - folder in diary could not be ascertained as the folders that file notes had been sent to had previously come out of diary. The report then gave a summary of results from performance as follows. A number of improvements were observed with the performance of the system from the test conducted on 1 June. These improvements included:

[383]

"Unisys' commitment under our agreement is to deliver a system which meets the Functional Specification. The purpose of the Acceptance Tests is to measure the system's compliance with those Functional Specifications. The Functional Specifications do not include any performance commitments. Performance levels are determined by a combination of the functional requirement and architectural constraints. In the case of RACV, the system currently being acceptance tested has achieved 'reasonable performance levels', in accordance with these factors. The Functional Specifications called for complete folders to be accessed, not single documents. The response times experienced are consistent with this approach and are also consistent with industry expectations, given the number of documents per folder. In our attempts to consider viable options for performance improvement in this area, we have been provided with varied opinions by RACV on the requirements of Claims Officers, regarding the depth of current and historical information to be made available at the desktop. We suggest that a detailed analysis be conducted by RACV to specify the precise functionality now required. We have, however, indicated the potential for Filenote retrieval improvements. We are surprised that you are now requiring outstanding performance criteria to be addressed prior to 24 June as it has previously been agreed that we would schedule the remaining four outstanding performance criteria to be addressed after implementation. We have also previously explained that performance of the system, as RACV currently propose to measure it, is outside the scope of what we are empowered to control in that we do not control the network, LAN, host etc. We have also, in a previous meeting, requested clarification from RACV on the accountability for 'end to end' performance. As yet, no assignment has been communicated to us. What RACV is now requesting is a fundamental change to what was originally requested and as such is an unreasonable expectation."

[397]

Counsel put the case in two ways, broadly speaking. The first was based on the proposal representations. The second way, which was counsel's primary way of putting the s. 52 case, might be called the configuration case. As I understand it, the case may be summarised as follows. Unisys knew that RACVI entered upon the WMS project with the aim of re-organising the claims handling process so as to produce more efficient and productive systems including document imaging that involved fewer staff and lower cost. Putting aside the complexity of re-organising the numerous processes involved, the system of electrical retrieval of images to the claims officers' screen, often while attending to a telephone inquiry concerning a claim, was intended to save time that the claims officer would otherwise spend in manually locating and retrieving the file and having to make a return call to the inquirer. To achieve this the computer imaging system had to be functional and stable. It also had to be efficient in terms of speedily retrieving images to the screen of the claims officer. Unless there was an appreciable improvement in this respect the business purpose of RACVI in entering into the contract with Unisys would not be met. To meet this business purpose the system had to be one which enabled images to be retrieved in a time consistent with an efficient on-line system. It happened that retrieval under such a system was in the vicinity of 2-4 seconds. Hence, it was submitted, Unisys had to configure the system as an on-line system for current claims. Timely image retrieval would follow as a consequence. Thus the primary way in which the case was put was on the basis of this configuration requirement rather than on the basis of a representation (or term of the contract) as to a retrieval or response time in the vicinity of 2-4 seconds. Notwithstanding its awareness of these matters Unisys configured a system in which current claims were partly on-line and partly near line. The result was that information sought by a claims officer could be contained on several disks, not necessarily then in page cache or in the claims officers' PC and not under a reader. The information could be spread over several disks which meant that retrieval of it could involve the spinning up and down of platters in a jukebox. The foreseeable consequence was response times of 20 seconds or so to minutes or more depending on the circumstances. The effect of response times of that magnitude on the ability to handle telephone inquiries without unreasonable delay, and the need to return calls, and otherwise for claims officers to work in an efficient way, is obvious. The system did not meet the business requirements of RACVI simply because Unisys adopted an inappropriate configuration. In seeking to establish a functional and stable system Unisys overlooked the fundamental requirement that information pertaining to current claims be retrieved more speedily and efficiently than under the existing system.

[408]

At an early stage of the trial, Unisys amended its defence to plead a reliance on a passage on p. 10-1 of the July response and point 12 on p. 7 of the PMP. The former is quoted at [93], and the latter at [138]. Reliance on these statements was pleaded at several points: in para. 14 as an answer to the plea of the proposal representations in para. 20 of the statement of claim; in para. 15A it was admitted that the intent and effect of the passage on p. 10-1 was that Unisys did not commit to any response times as part of the contract or otherwise (this admission related to the plea of misleading conduct by silence in para. 22(1) of the statement of claim); and it also relies on p. 10-1 and point 12 in para. 15D which pleads to para. 22(4) of the statement of claim. These paragraphs plead to the pre-contract s. 52 case. Then, in pleading to the allegation of a breach of contract it refers to p. 10-1 and point 12. The plea in para. 70 in relation to breach is more specific. After denying breach it is alleged in respect of the breach alleged in para. 36(f) and (g) (response times) that Unisys relies in particular on p. 10-1 and point 12 as providing that response times were to be finally negotiated once the system required by the functional specifications was installed. In summary, the defence relied on the passage on p. 10-1 and point 12 to establish that there was no representation or term as to response times from the system. The defence also stated that response times were to be negotiated once the system required by the functional specifications was installed. It is not clear whether that contention was alleged to be stated at either point or was to be implied or is in the nature of a contention. The reference in the defence to para. 36(f) and (g) was the only specific reference to the proposal representations and terms of contract alleged in the statement of claim. Paragraph 36(f) and (g) corresponds to para. 20(e) and (f); see the summary at [21]. The allegation in those paragraphs was that the system provide high speed image access for on-line claims with a retrieval time in the vicinity of 2-4 seconds and image access for near line claims with an expected response time in the vicinity of 20 seconds maximum.

[437]

At the appropriate point in the response Unisys stated the product it offered met the on-line and near line requirements in the RFP, which meant the configuration would provide the response time stipulated in the definitions. I reject the argument that the appropriate place to negative this commitment was at p. 10-1. That was some 64 pages on, on my count, and in a different section. Neither common sense nor a sense of fair dealing or reasonableness which one might have considered Unisys brought to the task of preparing the response, nor the wording of the RFP quoted at [76], required that the passage be placed at p. 10-1, assuming that in fact it was intended to exclude any commitment or preparedness to commit to the time in which the system would respond to the request of a user and bring an image to his or her screen. It was suggested by counsel for Unisys that the placing of the statement on p. 10-1 accorded with the requirement quoted at [76]. I do not agree. In the first place Unisys was required to deal with technical matters in Section Four. If the response was that it would not or could not commit to the on-line and near line requirements it would surely, acting reasonably and responsibly, have then and there said so. It beggars belief that Unisys would have offered the commitment in Section Four but sought to negative it in a later section much further on without a specific statement that that is what it was doing. Moreover, to so structure the response carried the risk that a reasonable reader might have been mislead. I would assess that risk as real and foreseeable. Indeed I note that Chesson-Kistas read the passage and understood it to relate to support; I note that but do not take that into account in determining objectively the point of construction. Nevertheless what she said, which I accept as her reaction on reading the passage, accords with what I would have regarded as the probable reaction of the reasonable reader. In my view, if the passage was truly intended to have the meaning contended for by Unisys its placement in Section Ten carried a real risk, indeed probability, of misleading and deceiving RACVI. The fact is that if the passage was intended to have and convey that meaning it would surely have appeared in appropriate terms in Section Four so as to clearly inform RACVI that Unisys would not or could not commit to the technical requirements of the on-line and near line system stipulated in the RFP. Further, if the passage was meant to have the effect contended for, it was not merely stating "additional information". Rather, it was negating the position stated in Section Four. So understood it seems incredible that Unisys could have intended to have indicated one thing in direct response and, much later, said the opposite in indirect response. In expressing these conclusions I do not overlook the various statements in the response mentioned earlier which in summary concerned the need for more definition of the requirements and to further consider the matter and that there may be changes. They do not meet the present point which is whether the passage on p. 10-1 excludes any commitment or preparedness to commit to response times which, in turn, takes one to the requirements in Section Four.

[438]

Counsel for the plaintiffs relied on some further factors as supporting this conclusion. The passage at p. 10-1 being general in its terms should not be read as derogating from the earlier specific positive representation relating to response times; see Forbes v Git [1922] 1AC 256, 259, and Joyce v Vaughan (1965) 66 SR (NSW) 156, 161 and 165. This is particularly so where the later clause is a standard form clause (Lewison, The Interpretation of Contracts, 2nd Ed. (1977) para. 6.04 and 8.08) which Josephson agreed it was. Then there is the principle that a general clause will not be interpreted so as to destroy the main object of the contract which here was concerned with improved efficiency in retrieving and handling files; see Margetson v Glynn [1892] 1QB 337, 334, and The Interpretation of Contracts, supra, para. 8.09. On its submission Unisys made no commitment to meet any retrieval requirement. Then there is the approach, particularly in a commercial context, that in the case of ambiguity of expression the parties are taken not to have intended an absurd or unreasonable result. Such a result would obtain if RACVI entered into the contract with no commitment from Unisys as to response times, and under which in terms of response times it would get whatever the system might come to provide. It is unlikely in my view that a potential vendor such as Unisys would have intended in its response to state that it would not commit to the stipulated response times, or, indeed, to any at all. If it had done so it would reasonably have expected that RACVI would not have further dealt with it, as it would in effect have been stating that it could not satisfy a critical element of RACVI's business purpose. While the RFP was not in itself a contract or deed between parties, it was nevertheless a document put forward in response to a request and Unisys intended it be relied on as a step to gaining a possible contract. Allowing for that, it is yet appropriate to bear these sort of considerations in mind in approaching the reading and construction of the July response.

[445]

The submission is readily understandable but it is curious in a way, just as the express non-inclusion of the RFP as a contract document may have been more unfortunate and productive of contention than wise or helpful. The position is this, commencing with (but not repeating) my observations at [140]. The contract was non-informative in that it did not state what the parties were setting out to achieve. The content or substance of what the system was to be able to do had to be found elsewhere. The contract documents which inform a reader of those matters are the July/October responses and that is perhaps why they were incorporated as contract documents, although that is speculation. But neither response is complete in itself. Each is responsive to and refers to the RFP. The responses cannot be fully understood or given full meaning without also reading and having regard to the contents of the RFP. It is thus correct in my opinion that even if the RFP does not possess the character of a contract document yet regard must be had to it to understand and give meaning to the responses. No amount of sophistry can avoid that consequence. It means that even if the RFP is not to be regarded as a contract document, it yet has effect in that way in explaining and giving content to the responses and hence understanding the nature of the task entered upon. The consequence is inescapable in my view that as a result of reading the responses with the RFP a contractual obligation may be identified or discerned with greater clarity than might otherwise have been the case. That is so, I consider, whether, strictly speaking, the RFP is considered as a contract document in itself as a result of necessary incorporation by reference in the responses (or perhaps even the PMP) or as a document to which necessary reference may be had to give content and meaning to a specified contract document otherwise uncertain, incomplete or lacking in understanding of the content or subject matter. The entire agreement clause does not stand in the way of that approach in this case.

[457]

The mention of the Olsen estimates raises again the issue as to when they were given to Doherty and by whom? There is no evidence on the issue. It may have been Birthisel or it may have been Esmonde who, according to Birthisel, was the person who was "managing that dialogue" with Doherty so that Birthisel could concentrate on the other things. In any event, it is conceded the estimates were provided to Doherty. At one point counsel for the plaintiff submitted that the apparent improvement involved in the Unisys estimated times might explain the suggestion that Doherty went along with the idea of performance enhancements after the system went live. As to why Doherty might have been so induced, counsel pointed out that for the first transaction type the estimate of 5-15 seconds took 10 seconds off the then minimum. If achieved, that would have been a very significant reduction. The submission draws attention to the total absence of evidence from Birthisel as to the alleged conversation, beyond the conclusionary assertions in his witness statement and oral evidence. It is left in the air. In effect, the plaintiffs' submissions suggested that the Olsen estimates were used to encourage Doherty to take a course in relation to response times that would defer that problem area until, hopefully, Unisys could succeed in having RACVI otherwise accept the system. Set in the context of the written communications and Hurford's attitude, and the evidence of Hurford, which I accept, that he had no knowledge of such an agreement, it is improbable that Doherty would have made the alleged agreement. I add that in the circumstance of the significant pressure he (and Unisys) was under knowing that its system could not provide the required response times and being desperately anxious (as I find he was) to avoid the obvious and looming possibility of the contract, which was already a loss situation for Unisys, being terminated, I find that Birthisel, and Esmonde, took every course possible to try to advantage Unisys in the circumstances. I readily find that they endeavoured to persuade Doherty to a position favourable to Unisys and that the Olsen estimates were provided accordingly. That does not mean, however, that the alleged agreement was made. It does, however, reflect a situation in which, as I find, Unisys would seek to create and fasten onto whatever point might be available and press that on RACVI in an attempt to protect and salvage its position.

[459]

The next document relied on was Walsh's letter to Hurford dated 23 May. Unisys relies on the reference to an agreement in the letter. I referred earlier to Hurford's evidence in which he accepted that an agreement had been made to defer work to enhance performance. Earlier in cross-examination he said, and I accept as correct, that he had no knowledge of such an agreement having been made before the contract was terminated. He was aware there were discussions about enhancements but not an agreement having been made as alleged. That means that Doherty did not advise him of such an agreement. His answer as to their having been an agreement came later. I am of the clearest view having observed Hurford give evidence that he was an honest witness and that neither answer reflected adversely on his credit. The question to which he assented was a general one which referred to a supposed agreement with respect to open-ended matters including "other work", "further enhance" and "consider". It was already clear he had no knowledge of any such agreement. I find he was in the dark and in giving the answer relied on he was going on the letter from Walsh. Further in his evidence, reflecting on the situation, Hurford said that he did not know that Unisys would be given a further opportunity to improve the system, but 24 June was getting closer, it was becoming apparent that not all of the timing requirements were going to be met, and obviously some consideration was being given to enhancing the system after 24 June. But whether or not there was "an agreement", by his letter Walsh stipulated the condition that what he called "this additional work" be "covered under new commercial arrangements". In other words, he wanted RACVI to bear all or part of the burden or cost of the work. RACVI never agreed to that. Hurford considered as ludicrous the notion that RACVI would pay Unisys for work which it was asked to do in the first place. Hurford's view was that there was a primary requirement on Unisys to deliver the system before any additional work was to be considered and performed. Hurford put it thus:

[479]

The document bears no signature or other signification of authorship or date of making. When a question arose as to whose document it was, the original was obtained by counsel for the plaintiffs and received in evidence as exhibit BB. The original established that the RACVI emblem appearing on the copy was only part of a sticky that had been attached by RACVI in the discovery process. The original document remained as one bearing no signification or date of authorship. The document in the court book was discovered by the plaintiffs and, it was said, from Doherty's files. But counsel for the plaintiffs said that it was not known who was the author of the note, and it was never admitted that any employee or agent of the plaintiffs was the author. The point was not revisited. No witness, including Doherty, was asked about the note when they gave evidence. I have noticed myself on reading the court book that the two numbered points in the first part of the note are word for word set out in Doherty's report on the 15 June test at p. 9936 of the court book. He there sets out the two points as matters indicated by an analysis of queue contents after the test. But the balance of the note as to Woozle transactions and the quoted passage concerning re-alignment of the optical disks is not set out in Doherty's report at p. 9936. It was not submitted that I should infer from the report at p. 9936 that Doherty was the author of exhibit BB. Nor do I think I can do that. I do not know that the platters were re-aligned by Doherty - indeed it is likely it was Olsen - or his involvement in or personal knowledge of what was done in that regard. Further, exhibit BB could have been prepared at some time subsequent to the Doherty report and by some other person on the basis of information provided by another person. In short, the authorship and the document is not established and nor is the date and circumstances of its making. The author of the document not having given evidence as to the note the contents remain a hearsay assertion. The note is not admissible, in particular as an admission by RACVI.

[498]

As I have said, the view of Dr Thorne that there was about one to nine gigabytes seemed to be accepted. He considered much more cache was required having regard to the volume of documents referred in the RFP. His view, and that of Hesom, was that about 12 gigabytes of page cache was required. On this basis the deficiency in cache was 10.5 gigabytes. Unisys' case is that that deficiency could have been resolved by the provision of an extra 10.5 gigabytes or so of magnetic storage space at a cost of about $40,000. It may be accepted that reforms to the claims processes would have resulted in the reduction in the number of images required to be held. The extent of the reduction was not calculated and would be speculation for me to venture an opinion upon. But whatever the reduction might have been found to be if the system had gone into production, it is a long stretch to suppose that it represented in the order of 10.5 gigabytes of magnetic storage space. It is safe to find, as I do, that the reduction was not of that order. At all times up to termination in June 1996 there was an observed slowness in response times which Unisys was not able to overcome. In all the circumstances, including the argument now made that the answer was to increase cache, the probability is that as at June 1996 cache was too small to hold all document images for current claims and I find that to have been the case. Reference may be made in this respect to evidence of Olsen that if, as I find to be the case, there was between one and two gigabytes of cache in June 1996, he believed it would not hold all current claims. In that regard too I note the concern expressed by McFarlane in October 1994 that there be enough cache to hold all current claims. The problem goes back to the fact the Unisys configured the system on the basis that cache was to be used as a transitory holding area rather than to hold all current claims; in effect as an archival system with current claims going near line. In any event, the result was that the system was never configured with sufficient cache to hold all current claim images on-line. To adopt the method now suggested, of adding cache to ensure that all current claim images could be held on-line, would have been possible but would have constituted a changed design philosophy from that adopted throughout by Unisys in applying the InfoImage product for RACVI's business purposes.

[543]

I am prepared to accept that if Unisys and RACVI, with inexhaustible patience and a continuing application of resources, had kept working away at the project at some time Unisys could have delivered a system which met the requirements of RACVI. I reject the case that this would have occurred within a matter of weeks or even within 60 days. I find that it would have taken longer on the balance of probabilities. It is important to remember in this respect that the preferred option of adding cache, the point at the forefront of Unisys' case in this respect, was not in the mind of Unisys at any relevant time. It was the litigation that brought it to Unisys' mind. I find that if RACVI had not terminated the contract when it did, and Unisys had just kept soldiering on as it was, it would not in a proximate sense have considered the option of adding cache and changing the system to one in which all current claim images were held on magnetic disk. However easy to implement that change might be, it was fundamental and it was not an option in the Unisys mind. Even if it be assumed that at some time Unisys would have considered it, and recommended its adoption to RACVI, I am by no means prepared to find that that would have occurred within 60 days of 24 or 27 June or thereabouts. There were other options, such as pre-fetching and a study that Esmonde had in mind, which would have taken time to investigate. Then there were issues of functionality that needed to be resolved. Then there were unresolved issues as to who would bear the cost of enhancing performance of the system. The approach of the advocate that all one had to do was add magnetic cache and, as though with the wave of a wand, all would be resolved, was simplistic and unreal in context, in my view. Moreover, however complex the task was, and the complexity of the project is not to be gainsaid, the fact is that it was to be performed in a commercial context in which performance might ordinarily be required in a timely way. That is the element which the Unisys submission fails to meet. It speaks of having established that 2-4 second response times for single image retrieval from cache would have been achieved had the system "ultimately" been installed. Of course, there was also the issue of retrieval of documents in a folder. To the question, but when would "ultimately" occur?, the answer I find is no more precise than within the next year but certainly beyond the next 60 days or so.

[558]

It would have been at odds with the position between the parties at senior level for an (unnamed) employee or employees of the plaintiffs to have so agreed. The making of such an agreement is improbable in the circumstances. Furthermore, if an agreement was made it was very quickly countermanded, in effect, between the principals in control of the transaction, although that is another matter as it is after the event. Then, to record something as having been "stated", without advice as to who stated it, and in what terms, and as to any response, is not necessarily to record a discussion between the parties in which they concluded an enforceable agreement to vary the live date of 24 June. It is consistent with the words used that a Unisys employee advised that Unisys could not meet the live date of 24 June and that 3 July was a more likely date to which it should be moved. Without evidence it is impossible to make a finding as to the tenor of the words or the level at which Unisys pitched the point. It is consistent with this being the approach of Unisys that there was reference to the lack of any contingency in the timeline but, even then, that sentence in the passage was not explained by evidence. It is also consistent with this understanding of the situation that the Unisys 9 May papers refer to 3 July 1996 as a forecast date and as an estimated date which Unisys hoped to meet. That is regarding the matter from Unisys' point of view. There is also a consistency with Birthisel having no memory as to the alleged change. In my view it is probable that he would have remembered the change if it had been agreed. In fact, he continued to have 24 June in his mind. It accords with the probabilities that Unisys, which knew it was in a tight and difficult situation, should try to get extra time but it does not follow that an attempt in that behalf was accepted in the sense of an agreement. I accept that Unisys did raise the issue and sought concurrence to a deferral but I do not accept that the quoted passage from Haw's memorandum, whether considered on its own or in the overall context, establishes that a variation agreement was made as submitted by counsel for Unisys or that as against the plaintiffs it constitutes an admission of the existence of the alleged agreement.

[580]

The first step is that a collection of documents known as a batch is scanned into the system in the mail room. The batch is sent to batch cache, part of a magnetic disk, in the server, to be indexed. Cache on the magnetic disk in the server is broken into four zones, batch, print, user and page. Batch cache receives the documents scanned in the mail room. Print cache contains documents requested by users that were in optical storage and being temporarily stored before being printed. User cache is very small, it stores information about user profiles and access information. Page cache is the balance and is available for document storage. Olsen said the documents in batch cache were indexed and copied to optical disk when an optical disk able to be written to was mounted. They were then flagged as belonging to page cache. This took seconds or minutes. The documents would remain in page cache until they were overwritten according to the rules of the system; that was when the system determined that additional cache space was required. At this point I add something to the account in the plaintiffs' submission. It appears that at different times pre and post-contract a calculation was made by Unisys (it seems by Olsen and Walker) as to when documents would move from page cache to optical disk. Olsen said he could not recall the detail of any calculation. Nor did he have a record of it. In his witness statement he described cache as a temporary holding area for electronic copies of images, the images being held in cache until it ran low of allocated space when documents which had not been accessed for some time would be sent from cache so as to clear space in the cache disk and allow new documents to be added. This was a general description and in that vein he said that the process was designed to occur frequently in which respect he said that if a document had not been viewed for say two weeks it was not likely to be held on the cache and more likely to be held only on the optical disk. In cross-examination he said cache was used for transitory storage. But, as mentioned, in this particular instance he could not recall the period of time calculated and applied in the design.

[589]

In arriving this conclusion I have considered a submission of Unisys that Olsen gave evidence that he left post-contractual details of configurations of the system at RACVI when the contract was terminated. It was said that RACVI had not discovered the information or given an explanation for failing to do so. Olsen did not use the expression "post-contractual details" in his evidence and there may be some doubt whether he was referring to post-contractual information. Whether or not he referred to pre or post-contractual information or both and whether it was contained in documents or on the RACVI PC's, I do not understand Unisys to have brought an application that RACVI make particular discovery of such information. The information was relevant to the case and was discoverable by the plaintiffs, assuming Olsen's evidence as to it is true. Having regard to the resources and effort applied to the case it is singular that no such application was made. Even apart from an application of that sort, Unisys could have arranged for Olsen and others to attend RACVI's premises to inspect the system it left behind and even to seek to locate the alleged information. It never did any such thing. It was said by counsel for Unisys, in explaining why Unisys had not inspected the system, that some change had been made to the system after termination, but that was not explored in detail to establish that it would have been futile to have made any such inspection. Unisys did not have its expert, or any person, attend RACVI to inspect the system to ascertain information from any point of view. If Olsen, for instance, had re-acquainted himself with the system and perused information he said he left behind he might be expected to have had a better memory on matters such as cache configuration than he said he did when he gave evidence. His lack of memory and non-production of any records enabled Unisys to make a submission based on the onus of proof. Further, Unisys did not call Walker, who was involved in making calculations, and gave no explanation for that failure. In my view Unisys sought to gain an advantage out of this state of evidence. I reject any submission of impropriety by the plaintiffs in not discovering the information allegedly left behind by Olsen. Indeed in all the circumstances and having had the benefit of observing Olsen, I am not satisfied on the balance of probabilities as to just what he left behind. Further, is it to be assumed that what he left was in a form or manner of presentation that ought to have been recognised as containing what he claims? I conclude this section with a reflection on Olsen's claimed lack of recall of the design of cache and his sensitivity to admitting the fact, as I find it to be, that Unisys did not design the system to hold all current claims in cache. The former evidence strained credulity in my view, and I do not accept it. The latter was indicative of an awareness of a point embarrassing to Unisys.

[595]

If any witness for Unisys was to state the configuration in 1993 and subsequent to the contract it was Olsen, but he did not do so. He could not remember, he said, and we then had Unisys seeking to establish that the onus of proof of the configuration was on the plaintiffs. To refer to a few more points in Olsen's evidence, he did not know how much cache was required for active claims, he did not recall having made any calculations of the amount of cache required between January 1994 and March 1995, and he doubted that the system supplied in March 1995 or as it was in June 1996 was configured to hold all active claims in cache. His doubt, as I have found, was correct. I refer to another piece of Olsen's evidence. He had said that when designing the system he had a performance target in mind of 2-4 seconds retrieval time as in the RFP and 20 seconds for near line. In other words, he was aware of the statements in the RFP. But, when asked whether he configured the system so there was a reasonable expectation of meeting the target, he said "We didn't look as closely". I find that reflected the true situation. Unisys just applied its InfoImage product with unmounted jukeboxes instead of an on-line system. Andrew Walker worked with Olsen in relation to the configuration of cache but Unisys did not call him. I accept the plaintiffs' submission that if called his evidence would not have assisted the Unisys case in this regard. Birthisel never considered configuring the system so that all active claims were on-line. He understood, when he read the RFP in March 1996, that the RFP required all current claims to be on-line. But, much later in cross-examination, his evidence was that it was never intended by Unisys that the optical disks storing current claims would be in a fixed position mounted on their own readers. When he saw the section in the RFP about a 2-4 seconds response time, the focus was on getting the best performance for RACV rather than meeting that target. He said that from 1994 to 1996 the system was configured so cache was used as a transitory holding area rather than to retain all active claims, and that Unisys intended to hand it over in June 1996 in this state. It was an archival system with claims held near line rather than on-line. Josephson could not say how much disk space was allowed for cache in the originally proposed configuration, or subsequently.

[623]

Unisys sought to draw a point out of the 29 September letter. In relation to final costings the letter requested a firm and final cost on, inter alia, "Configuration (both centralised and distributed - as per attached)". The attached sheets diagrammatically depicted a centralised and distributed image system LAN layout. The Unisys October response answered the request; see at [124] above. Unisys submitted that it was free to make whatever configuration it considered appropriate, and that response times were not raised as an issue. It was submitted that the letter was evidence that no person at RACVI understood or represented any particular configuration at all. Of course the submission was made in the context of the reliance by Unisys on the qualification on p. 10-1, which I have dealt with. In my view the Unisys submission is without substance. All that RACVI's letter did was request a "final submission for provision of our Workflow/Imaging system". That was a reference to the system the subject of the RFP to which Unisys had provided a response in July followed by a demonstration. All that RACVI was now doing was providing Unisys - which had sought it - the opportunity to provide a final submission. But the submission nevertheless related to the subject matter dealt with, and the requirement stated in, the RFP. Further, the 29 September letter said nothing about the on-line requirement stipulated in the RFP and no advice or withdrawal or variation of that stipulation had been provided to Unisys by the plaintiffs. There was therefore no need to restate the requirement. It is obvious from the brevity of its terms that the letter is premised on the RFP. Moreover, the reference to configuration in the letter was not to the particular aspect of the architecture or design of the system that would result in current claims being stored on-line as stipulated in the RFP. It was configuration in an overall sense of centralised versus distributed and recommended, and costings, that the letter was concerned with.

[635]

In their submissions the plaintiffs said this analysis was incorrect, and made reference to several passages in the evidence that indicate the actual thinking and approach of Unisys was to the contrary. There was the evidence of Josephson as to the demonstrations which was consistent with the requirement in the RFP and the Deloittes 1992 study. And Josephson's evidence constituted an acknowledgment that retrieval times were an important objective. They were important from the point of view of configuring the jukeboxes to provide the required retrieval speed. Schmura said the issue of retrieval times was raised with him within Unisys by Bruce Wilson (who was not called by Unisys) in early 1995 and Rouqeuriol in June 1995. The August meeting in the United States referred to folder retrieval to a desk top taking several minutes as a major performance issue. Johnson agreed that in August 1995 slow retrieval times were perceived as carrying a lot of problems. Olsen agreed a retrieval target of 2-4 seconds in some of the early demonstrations was an important factor. He regarded it as an achievable target when he was on the design team, and he had that view in July 1993. He also said that a purpose of introducing the single server was to enhance performance, of which one aspect was response times. Then, in an internal Unisys memorandum dated 18 October 1993 it is noted under the heading Major Problems that inaccurate sizing of the system will require upgrade. Then, moving to 1996, there are the letters from Walsh which referred to "the performance criteria originally specified in the contract" (letter 7 March) and "the performance criteria set out in our initial agreement" (letter 23 May). No performance criteria was specified or set out in the contract. The references must be, and I find they were, to the requirements of an on-line system specified in the RFP. The fact is, I find, that a fundamental premise of the engagement, understood by Unisys, was that the system implemented would provide retrieval in a timely manner. So much was obvious as required to meet the business purposes of RACVI. But over and above that was the requirement of an on-line system in which the retrieval times specified in the RFP were inherent. This was understood by Unisys and hence the explanations and representations of Josephson and Unisys' cognisance of the importance of retrieval time signified by the above references. I reject the submission that under the contract Unisys was to deliver, and RACV was to accept, a system with whatever retrieval times it might come to deliver.

[640]

I find that if RACVI had understood the position of Unisys to be that it was not committing itself to provide a system that met the requirements of these representations it would not have dealt with Unisys. It would not have entered into the contract. The statement of the mandatory requirement in the RFP was clear and understandable as fundamental to a system to meet RACVI's need for a more efficient system. In direct terms in the July response Unisys stated it would provide a system which "allows for the on-line and near-line storage requirements outlined in the RFP", and that Unisys had "configured a system which can comfortably handle the committal and retrieval rates expected from the volumes outlined in the RFP [and RFI]". Nothing could have been clearer. It seems a singular way of conducting business to then bring in aid some general statements as to the need for more precise information (which in itself was understandable) and a supposed qualification in a different section to negate the commitment. If that was truly the way in which Unisys approached the matter, to say one thing in direct response to induce the other party to have comfort in dealing with you, and at the same time, but without then and there squarely saying so, you use words elsewhere to give you the option of setting the system up how you like and contrary to the earlier commitment, then the seeds of contention were sewn early. But that is not how the plaintiffs read the RFP and, in my view, that is not how it was reasonably to be read and understood. It seems difficult to countenance that Unisys intended by the qualification on p. 10-1 to negate the earlier direct statement, and I have concluded against that construction. But if that was the intention there is room for the inference in my view that it was positioned where it was, up the back of the document and far removed from the earlier section, and obliquely expressed, in order that it not be seen for what it was and where it would not be expected. The misleading or deceptive nature of such a course is apparent. Further, in my view the other general words elsewhere in the July response were not reasonably to be understood, whether considered alone or in context with other Unisys statements, as negating the statements as to configuration as being statements of no intent or on which no reliance could be placed.

[687]

I turn then to the facts of the case. First, there were the facts of the relationship between the plaintiffs referred to at [3]. Then there was the working relationship between the plaintiffs and between them and Unisys. RACVGS staff were involved both pre and post-contract in working on the development of the WMS project, as Unisys knew. Unisys also knew and understood that RACVI and RACVGS were separate but related entities under the RACV umbrella and that RACVGS was responsible for computer services for RACVI. RACVGS was even a party to the PMP which itself reflected RACVGS' involvement in the contract and its implementation and Unisys's knowledge of that involvement. Indeed during the trial Unisys placed reliance on the knowledge and expertise of RACVGS in relation to various matters as a fact establishing the knowledge and understanding of RACVI on relevant matters. It is evident that Unisys approached the matter from the point of view of dealing with RACV and did not pre or post-contract draw fine distinctions between the roles and work of those who strictly speaking were employees of RACVI on the one hand and employees of RACVGS on the other hand. In short, to the knowledge of Unisys, RACVGS had an active role as an RACV company associated with RACVI in the development of the WMS project and the implementation of the contract and in those matters it dealt with Unisys. It was in that context that RACVGS sent cheques to Unisys in payment of certain RACVI invoices. I find that it did so, and that it was induced to do so, by the representations I have found to have been made, of each of which RACVGS was then and there aware by reason of its involvement. I find that Unisys was aware that RACVGS acted in reliance upon it and in particular the representations. The making of the payments also followed from the fact of the relationship between RACVI and RACVGS and as a result of RACVI having been induced to select and contract with Unisys. In my view the payments directly resulted from Unisys' conduct in contravention of s. 52.

[689]

These submissions accurately noted the actual words of the pleading. There are some points which should be immediately noted. First, the submission gives to the words "pursuant to the WMS contract" in para. 29B a literal reading which they cannot bear, and were obviously not intended to bear, when understood in light of the allegations in the several sub-paras. in para. 29B. Many of the things there referred to as having been done by the plaintiffs were not required to be done by the terms of the contract, as distinct from being something done following or consequent upon the contract, to note some meanings of the word "pursuant" in the Oxford Dictionary. Secondly, when the claim of a contract with RACVGS in paras. 29 and 29A was abandoned the statement of claim could have been, and still could be, amended if it were necessary in the interests of justice to do so, that is to say, to ensure that the matters truly in issue were raised for determination. Thirdly, once the abandonment was made, and RACVGS' s. 52 claim was still pressed, it was obvious that the expression "pursuant to the WMS contract" in para. 29B was to be read from the point of view of RACVGS as meaning "as a result of" or "because of". The sense of the thing, and the way in which RACVGS puts its claim, is obvious. So also was it obvious from an early stage of the trial, and it must have been to Unisys who at all times had denied that RACVGS was a contracting party, that the allegation would not succeed. The ultimate concession reflected the obvious; on the evidence the contract claim was hopeless. I record these matters to ensure a reality check in respect of Unisys' submission. In no sense could Unisys have been misled by the terms of the pleading in question to misapprehend the position it had to meet and as it was finally conceded, and in no way has the pleading and the pressing of the s. 52 claim following the abandonment of the contract claim caused any prejudice to Unisys in the presentation of its case. Rather, according to Unisys, RACVGS "has limited its case in a manner not required by the terms of the statute, but it cannot escape the consequences of its own pleading". It is a pleading point, and no more. It would be a denial of justice, indeed in the circumstances to the point of being a travesty, to uphold the submission.

[697]

These submissions were opposed by Unisys at every step in the argument. The immediate issue is whether the amount of $84,070 was in respect of a claim that was separate from the common or overlapping claims against the defendants. In my opinion it was not attributable to a separate claim. In identifying whether a claim is separate and additional to a common or overlapping claim the issue is not whether the claims are based on the same cause of action but, rather, "whether the claims against the various defendants are 'concurrent' in the sense that the relief sought is the same" (per Winneke P in Bonchristanio at 89). In the present case the loss and damage alleged to have been suffered by the plaintiffs and in respect of which damages are sought in the statement of claim, was identical. This has been clear since the statement of claim filed pursuant to O.14 on 30 June 1997 in which the prayer for relief claimed from Unisys and Deloittes the same items of loss, including the same amounts paid to Deloittes. The amount claimed as paid to Deloittes was stated to be $381,953. The same prayer for relief was carried forward in the amended statement of claim filed on 5 June 1998. In its subsequently amended forms the statement of claim omitted the prayer for relief, it would seem by oversight, but in its final form the loss and damage claimed to have been suffered as a result of the breaches or conduct of the defendants was stated to be set out in annexure B to the statement of claim. In the section of annexure B dealing with the amounts paid to Deloittes, the amounts were itemised and produced the total of $422,615. Quite plainly, that meant that the plaintiffs were claiming that amount from each defendant. Neither in the statement of claim nor in annexure B was a distinction drawn between the amount claimed from the defendants in respect of any of the items of loss and damage. The damages claimed against each defendant were identical. It follows that the claims against the defendants were common or overlapping and that there was no separate or additional claim against Deloittes in respect of the pre-contract payment of $84,070.

[700]

In answer, counsel for RACVI acknowledged the difficulty, including that the costs of the claim against Deloittes would have been "intermingled" with the costs of the proceeding against Unisys. He submitted that the present situation was "an instance of the Court doing the best it can, from its own knowledge of the expense of this type of litigation, to estimate the fees referable to the claim against Deloittes". There had been extensive pleading and it was submitted that I could reasonably estimate the costs at $200,000. I have perused the court file which runs to five folders and records numerous pre-trial interlocutory hearings. It is apparent that the case has occasioned a significant degree of contest. It was a complex case which would have involved considerable time and attention to the facts and discovery. In making these observations I have the benefit of having been the trial judge and thus derived much insight into the dispute, including having heard MacCallum give evidence, and an appreciation of the considerable burden of work that would have been involved for the practitioners. The settlement with Deloittes was not reached until August 2000, which was only several months short of four years of litigation. It was said that $200,000 was a very conservative sum, and that observation may be correct. My difficulty is that I do not possess the experience and knowledge of a Taxing Master and in the absence of evidence am unable to make a reliable assessment of the costs. What, then, is to be done? I do not disregard the fact that RACVI could have sought to establish by evidence the amount of the costs, and did not do so. It may be that the complexity and cost of the litigation was a factor in not adducing such evidence. But does the absence of probative evidence mean that RACVI's conceded entitlement to an allowance is to be denied for want of sufficient proof at this stage? In my view, in the circumstances of this case, the interests of justice are best served by orders which provide for a reference to the Taxing Master for the assessment by him of RACVI's party and party costs of the proceeding against Deloittes, the sum so assessed to be allowed against the settlement sum. The amount, if any, thus ascertained as the balance remaining of the settlement sum will then be allowed in reduction of any damages payable by Unisys in respect of the amount paid to Deloittes or of the damages generally (it is not necessary to decide which is the correct method although the latter would seem correct). As mentioned, RACVI cannot recover the amount of $84,070 paid pre-contract to Deloittes. Hence, as mentioned above, the amount of the Deloittes fee recoverable from Unisys is $338,545 and it is against that amount (or the damages generally) that any excess of the settlement sum is to be allowed.

[737]

At the first point, there was an objection concerning Blashki's evidence. In the discussion which followed I inquired as to the use to be made of the evidence. At first counsel said, consistently with his opening, that the evidence was produced "effectively to support" the claim for wasted staff costs put forward by Doherty (who had given evidence the previous day and to which evidence I refer below) "and that's all". He said, correctly, that the calculations of Blashki and Doherty were done on a different basis and that if the amount calculated by Blashki was claimed separately "it would be a duplicitous claim". (It would be duplicitous because the labour costs which formed the basis of Blashki's calculation included the costs of the particular employees which the plaintiffs sought to recover under their claim for wasted staff costs which in part Doherty was called to advance.) But, in response to a further question, counsel changed his ground. He said that if the claim for wasted staff costs calculated by Doherty was not accepted, RACVI sought the amount calculated by Blashki. As finally presented by Blashki that amount was $654,874. I had the clear impression at this stage of the case that counsel was uncertain and somewhat feeling his way as to what would become of the claim and how and to what extent it might come to be presented in final address. In final address counsel changed his ground on the Blashki aspect. (He also changed ground on the matter of Doherty's evidence but I refer to that below.) He did not submit that RACVI was entitled to recover the $654,874 calculated by Blashki. Nor did he submit that the Blashki calculation was there to merely support the primary claim of wasted staff costs. Rather, and, it would seem, on the basis of accepting Blashki's evidence but doing a simpler exercise to establish the labour cost per claim increased in the periods March 1994 to February 1995 and March 1995 to February 1996, it was submitted that I could conclude that a figure "substantially less" than Blashki's figures "could be estimated by your Honour as a measure of the wasted labour costs as a result of the reduction in productivity during the time it was sought to implement the system". The figure put forward as that which could be estimated was $250,000. That is the amount claimed and that is how it was arrived at. It was not Blashki's figure. No witness, expert or otherwise, put it forward as a calculated estimate of loss, and it was not put forward as calculable from evidence given. It is a figure in the air put forward as a minimum or conservative estimate which I should feel comfortable in adopting.

[739]

The second point is that having carefully considered the evidence I am not satisfied, on the balance of probabilities, that it does establish what Blashki concluded. That is because, in my view, the evidence is not reliable. Blashki assumed that the wage increases in the period in question were those advised to him by RACVI; and see the evidence of Peter Malcolm Jenkins as to those rates. A critical step in Blashki's calculation was to represent those wage increases as a percentage annual salary rate increase in the two periods in question. In cross-examination of Jenkins, based on actual wage records of RACVI, it was established that there were wage increases at appreciably higher percentage rates than advised to Blashki. Blashki could only say that he had accepted as correct the wage increases advised to him by RACVI. I find that the increases advised to Blashki were changes to the base salary, and not overall wage increases in fact. Furthermore, the figures provided to Blashki were not for all the workforce in the relevant departments. RACVI had no sound answer to the resultant criticism of the inadequacy of the information on which Blashki acted. It was not an answer to talk about the changing profile of a workforce with new employees coming in at the bottom and others going out at the top. Blashki was concerned with overall wages in the two RACVI departments in question and from them had to determine an annual salary rate increase. To only consider the changes in the base salary to part of the workforce without taking account of other movements in salary for all of the relevant workforce was to take account of part but not the whole of the relevant information. In cross-examination Blashki agreed that if the annual percentage rates of increase were those put to him by counsel for UNISYS his calculations would alter to the extent that on the logic of Blashki's exercise, RACVI not only had no additional labour costs as a result of the WMS project but had saved wages. The full extent of this saving was subsequently calculated and in final address put by counsel for Unisys at $240,000. It is important to note that the percentages put to Blashki by UNISYS's counsel were not challenged by RACVI. They were stated by UNISYS's counsel to have been calculated from RACVI's wages records, and I accept that as the fact. RACVI had ample opportunity to go to the records and demonstrate that UNISYS had misused them but RACVI never suggested that Blashki or myself had been misled with inaccurate information in this respect. To conclude this paragraph, whether it is the fact that RACVI saved an amount of about $240,000 would involve an analysis of its business of a more profound and rigorous kind than was provided to me. Nevertheless the exercise engaged in by counsel was sufficient to demonstrate the unreliability of the Blashki exercise.

At a glance

Court

Decision date

Before

Source

Judgment (775 paragraphs)

Parties

Cases Cited (4)