A-332-01
2003 FCA 30
Correctional Service of Canada and Commissioner of Corrections (Appellants)
v.
Matthew G. Yeager (Respondent)
Indexed as: Yeager v. Canada (Correctional Service) (C.A.)
Court of Appeal, Stone, Isaac and Malone JJ.A.-- Ottawa, May 22, 2002 and January 22, 2003.
Access to Information -- Criminologist and penal system critic requesting from Correctional Service of Canada (CSC) data, code book, software -- CSC Research Branch conducting statistical research on corrections -- Having access to classified databases of other government institutions -- Information accessed by electronic links to computer networks -- Requested data not existing but could be recreated -- Requiring considerable resources, expertise over extended time -- Necessity for purging data of personal identifiers -- Denial of requests upheld by Information Commissioner -- F.C.T.D. Judge ordering production of data, code book but not software as for use only on mainframe computer, not "record" as defined by Act -- Novel issues before F.C.A.: is software a "record", must government institution create, furnish non-existent records -- Data, code book meeting definition of "record" in Access to Information Act, s. 3 -- Under s. 4(3), record that does not exist deemed record if can be produced from machine readable record under control of government institution -- Appellants' argument, statute not requiring creation of something not existing in any form, rejected as contrary to wording of provision, statutory construction principles, purpose of legislation -- Respondent's request was for reasons approved of by S.C.C. in Dagg case -- Limitations on disclosure obligation -- F.C.T.D. Judge ignored significant evidence, erred in concluding CSC operations not unreasonably interfered with if having to disclose information requested -- Difference between data, code book one of degree: disclosure of latter requiring more effort -- Going to unreasonableness issue -- Nothing turning on fact software for use with mainframe -- Real issue: is computer software existing "record" within Act, s. 3 -- Software not "record" but used to generate, view, edit record -- Software not analogous to items listed in s. 3 -- Under external developer's control, not government institution's -- Costs awarded unsuccessful respondent as appeal raising important new principles.
Construction of Statutes -- Whether certain information government institution refusing to disclose "record" within Access to Information Act, s. 3 -- Whether s. 4(3) (requiring production of certain material that does not exist but can be produced from machine readable record) to be interpreted as requiring creation of something not existing in any form -- Interpretation suggested by institution offending numerous statutory construction principles: "plausible meaning rule", "presumption of linguistic competence", "rule against tautology".
Practice -- Judgments and Orders -- Reversal or Variation -- Motion for reconsideration under Federal Court Rules,1998, r. 397(1)(a) -- Order issued by Motions Judge upon judicial review of government institution's denial of access to information request not dealing with all issues in reasons -- Problem discovered only when reasons for judgment in preparation by F.C.A. -- Further hearing held, submissions received from counsel who accepted Court's suggestion: motion for reconsideration as order not in accord with reasons and amendment of appeal pleadings.
Practice -- Costs -- Motions Judge making no order as to costs as success divided upon judicial review of denial of access to information request -- Requester unsuccessful before F.C.A. but awarded costs throughout, on party-and-party basis, as appeal raising important new principle in relation to legislation -- Dissenting Judge would have awarded $20,000 lump sum under Federal Court Rules, 1998 r. 400(4) as best reflecting Parliament's intention reasonable costs be awarded.
This was an appeal and cross-appeal from the order of a Motions Judge upon respondent's application for judicial review of a decision made by the Correctional Service of Canada (CSC) and the Commissioner of Corrections.
The Research Branch of CSC conducts statistical research in the corrections field. It is granted access to classified databases maintained by such agencies as the RCMP, Canadian Police Information Centre and the National Parole Board. These databases contain sensitive information on victims and offenders. Access thereto is restricted to those with enhanced security clearance. The CSC accesses this information online using electronic links to the agencies' computer networks. Matthew Yeager, the respondent, is a criminologist and a critic of the Canadian penal system as well as the policies that underlie the system. Yeager requested that appellants furnish certain information related to this research: data, a code book needed to interpret the data and software enabling use of the data on a personal computer. The data was the CSC release cohort used to recalibrate the General Statistical Indicator of Recidivism (with personal identifiers deleted). The code book used to define and identify the variables and a copy of the Offender Intake Assessment software were also requested. The evidence was that the requested data did not exist but could be recreated. This would involve doing the research project all over again and would require the dedication of considerable resources and expertise over an extended period of time. Additional effort would be required to purge the data of all personal identifiers. As to the request for a code book, such had not been prepared for the data requested. The software requested by respondent exists in a format for use on a mainframe computer system rather than on a personal computer (p.c.). This software was a massive repository of text files, the development of which had been a multi-million dollar project. Appellants rejected all of Yeager's requests, taking the position that there was no statutory requirement to create records that do not exist. Following an unsuccessful complaint to the Information Commissioner, respondent applied to the Trial Division for judicial review.
The Motions Judge held in Yeager's favour, concluding that CSC had failed to discharge the burden of establishing that production of the data and code book would unreasonably interfere with its operations. The Judge did conclude that the software does not exist in p.c. format and that, in any event, it is not a "record" as defined by the Access to Information Act.
It was only when the reasons for judgment of this Court were being prepared that it was noticed that the order issued by the Motions Judge failed to deal with all of the issues covered in her reasons for order. A further hearing was accordingly held at which submissions were received from counsel as to how best to proceed. Counsel accepted the Court's suggestion that respondent move before the Motions Judge, under paragraph 397(1)(a) of the Federal Court Rules, 1998, for reconsideration of the order as not in accord with the reasons, and that pleadings in the appeal be amended accordingly. This was done and the Court concluded that the cross-appeal had been properly constituted.
The Court was, for the first time, required by this appeal to interpret subsection 4(3) of the Access to Information Act and section 3 of the Access to Information Regulations in determining whether computer software is a "record" and whether appellants have to create and furnish records that do not exist.
Held (Malone J.A. dissenting in part), the appeal should be allowed, the cross-appeal dismissed but respondent should be awarded costs both here and below.
Per Isaac J.A. (Stone J.A. concurring): Following the teaching of the Supreme Court of Canada in Dagg v. Canada (Minister of Finance), the issue for resolution herein was whether each of the data, code book and software is a "record" under CSC's control. Under the definition of "record" found in section 3 of the Act, the data and code book would be a "record", but neither exists. The data requested would have to be recreated and modified and a code book prepared. While the software does exist, it is less clear whether it is a "record".
Under subsection 4(3), "any record . . . that does not exist but can, subject to such limitations as may be prescribed by regulation, be produced from a machine readable record under the control of a government institution . . . shall be deemed to be a record". Appellants' argument was that this statutory provision should not be interpreted as requiring the creation of something that does not exist in any form. They insist that there is no obligation to create a record that does not already exist in machine readable form. That interpretation is in stark contrast with the wording of the provision, contrary to the principles of statutory construction and went against the overarching purpose of the access to information legislation. Subsection 4(3), which applies where a record "does not exist", does not contain a clause "other than in machine readable form" or words to that effect. Since subsection 4(1) contemplates that a machine readable record must be supplied, subsection 4(3) would be devoid of meaning if interpreted as suggested by appellants. The principles of statutory construction offended by appellants' interpretation include: the "plausible meaning rule", the "presumption of linguistic competence" and the "rule against tautology". In enacting subsection 4(3) Parliament must have contemplated two different records: a new and distinct record must be produced from an existing machine readable record. Respondent's access request is for the very reasons contemplated by the Supreme Court in Dagg: to participate in and ensure the accountability of Canadian government research and to analyse the Canadian penal system.
It remained, however, to determine the meaning of the phrase "from a machine readable record". The answer to that was contextual and fact-specific. Whether a record is producible "from" a machine readable record depends upon several factors, including the amount of independent composition in comparison with purely mechanical editing.
There are, however, limitations on the obligation on government institutions to produce otherwise non-existent records. Subsection 4(3) states that non-existent records have to be produced only where capable of production "using computer hardware and software and technical expertise normally used by the government institution". Then, under section 3 of the Regulations, a non-existent record "need not be produced where the production thereof would unreasonably interfere with the operation of the institution". The Motions Judge ordered provision of the data and code book having found that appellants had adduced no evidence of unreasonable interference. In so concluding, the Judge below ignored significant evidence in the record. On the evidence, it had to be concluded that compliance with respondent's request would constitute unreasonable interference with appellants' operations. In particular, the Motions Judge wrote that she had "no idea whether it has a large staff and a complex computer system . . . [or] a small staff and limited computer capacity". There was, in fact, testimony that CSC had many databases, 1200 workers and many researchers. There was on the record an abundance of information regarding CSC's computer and human resources. Nor did the evidence support the Motion Judge's comment that she had been given no indication as to whether any important or urgent matters were being researched. Rather than examining all the evidence presented, the Motions Judge was extremely selective in the evidence she considered. In addition, the Judge placed too high a burden upon appellants to justify their refusal. The Supreme Court judgment in Reza v. Canada was authority for the proposition that, where a motions judge has failed to give sufficient weight to all relevant considerations, an appellate court is entitled to intervene.
Appellants failed in their attempt to distinguish between the data and the code book: the difference between the two was merely one of degree. Production of a code book would require somewhat more time and effort, but would nevertheless be prepared "from" the underlying machine readable record. This greater effort does, of course, impact upon the issue of unreasonable interference with the appellants' operations. Again, the evidence did demonstrate that producing a code book would unreasonably interfere with appellants' operations.
As for the cross-appeal, respondent's argument was that the Motions Judge erred in assuming he had to have the software in p.c. useable format. That submission was not without merit and the essential question was as to whether the mainframe computer software is an existing "record" within Act, section 3. The Motions Judge did not err in concluding that the software was not a "record". Software is used to generate, view or edit a record; it is not a record itself software is not analogous to any of the items listed in section 3. Furthermore, the software is not "under the control of a government institution"--it is under an external developer's control, appellants being mere licensees.
As held by the Motions Judge, there was here no infringement of respondent's Charter right to freedom of expression.
Turning to the matter of costs, the Motions Judge made no order as to costs, in that "success was evenly divided". The appeal and cross-appeal, however, required this Court to, for the first time, consider Act, subsection 4(3) and Regulations, section 3. Under section 53 of the Act, the Court may award costs to applicant, even if unsuccessful, if of opinion that the review application has raised an important new principle in relation to the Act. The issues raised herein do involve important new principles and respondent should be awarded his costs throughout on a party-and-party basis.
Per Malone J.A. (dissenting in part): A lump sum award under subsection 400(4) of the Federal Court Rules, 1998 would best reflect Parliament's intention that respondent be awarded his reasonable costs rather than party-and-party costs under the Tariff. A lump sum of $20,000 should be here awarded.
statutes and regulations judicially
considered
Access to Information Act, R.S.C., 1985, c. A-1, ss. 2(1), 3, 4(1)(a),(b),(3), 12(1), 41, 53(1),(2).
| Access to Information Regulations, SOR/83-507, s. 3. |
| Canadian Charter of Rights and Freedoms, being Part I of the Constitution Act, 1982, Schedule B, Canada Act 1982, 1982, c. 11 (U.K.) [R.S.C., 1985, Appendix II, No. 44], s. 2(b). |
| Federal Court Rules, 1998, SOR/98-106, rr. 397(1)(a), 400(4). |
cases judicially considered
followed:
Dagg v. Canada (Minister of Finance), [1997] 2 S.C.R. 403; (1997), 148 D.L.R. (4th) 385; 46 Admin. L.R. (2d) 155; 213 N.R. 161; Housen v. Nikolaisen, [2002] 2 S.C.R. 235; (2002), 211 D.L.R. (4th) 577; [2002] 7 W.W.R. 1; 219 Sask. R. 1; 10 C.C.L.T. (3d) 157; 30 M.P.L.R. (3d) 1; 286 N.R. 1 (S.C.C.); Reza v. Canada, [1994] 2 S.C.R. 394; (1994), 116 D.L.R. (4th) 61; 21 C.R.R. (2d) 236; 24 Imm. L.R. (2d) 117; 167 N.R. 282; 72 O.A.C. 348.
considered:
Ontario (Attorney General) v. Fineberg (1994), 19 O.R. (3d) 197; 116 D.L.R. (4th) 498; 25 Admin. L.R. (2d) 123; 73 O.A.C. 311 (Div. Ct.).
authors cited
Sullivan, Ruth. Driedger on the Construction of Statutes, 3rd ed. Toronto: Butterworths, 1994.
APPEAL and CROSS-APPEAL from an order of a Motions Judge ((2001), 204 F.T.R. 297) upon an application for judicial review of a Correctional Service of Canada decision refusing disclosure of certain information. The Judge below held in requester's favour with respect to two of the three items requested: data and a code book, and against him as to software. Appeal allowed, cross-appeal dismissed but costs awarded respondent throughout on a party-and-party basis.
appearances:
Christopher M. Rupar for appellants.
Neil R. Wilson for respondent.
solicitors of record:
Deputy Attorney General of Canada for appellants.
Gowling Lafleur Henderson LLP, Ottawa, for respondent.
The following are the reasons for judgment rendered in English by
[1]Isaac J.A.: This is an appeal and cross-appeal from an order made by a Motions Judge in the Trial Division, dated 3 May 2001 (reported as (2001), 204 F.T.R. 297) and one made by the same Judge nunc pro tunc on 20 January 2003. The order was made upon an application by the respondent for judicial review of a decision of Correctional Service of Canada and the Commissioner of Corrections (collectively, CSC).
[2]In his amended originating notice of motion dated 26 June 2000, the respondent sought the following relief:
(a) a declaration that the decision of the Respondents Correctional Service of Canada and Commissioner of Corrections ("CSC") contravenes the Applicant's constitutional rights pursuant to section 2(b) of the Charter of Rights and Freedoms, 1982;
(b) a review of the decision by the Respondents CSC to deny the Applicant access to:
(i) a copy of the Offender Intake Assessment software (current version in operation), which includes the: Custody Rating Scale (CRS), the CSIR, [sic] and the Community Risk/Needs Management Scale, among other features; and
(ii) the following micro-data:
(a) the 1993-93 [sic] CSC Release Cohort currently being used to recallibrate [sic] the GSIR. The micro-data is in the form of individual cases in a rectangular, fixed format. Personal identifiers should be deleted (such as name of inmate/parolee, FPS number, or full date-of-birth, however year of birth will not violate privacy).
(b) the Code Book used to define and identify/locate the variables in each case.
and an Order compelling the Respondents CSC to provide the Applicant with the foregoing information as requested under the Access to Information Act;
(c) the Applicant's costs of this Application on a solicitor-client basis; . . . [Emphasis added.]
[3]Although in reasons for order the learned Motions Judge addressed each head of relief sought, the first order in appeal reads:
NOW THEREFORE THIS COURT ORDERS THAT, for the reasons issued this day:
| 1. The Respondents are to supply the Applicant with the 1992-93 CSC release cohort used to recalibrate the General Statistical Indicator of Recidivism. The data is to be edited to delete all personal identifiers; |
| 2. The Respondents are to supply the Applicant with a Code Book which allows him to interpret the data produced under paragraph 1. |
Accordingly, the order fails to deal with the five issues dealt with in the reasons.
[4]By notice of appeal, dated 31 May 2000, the appellants ask that the order of the Motions Judge be set aside. Although the order in appeal is limited to the two grounds raised by the appellants in their notice of appeal, the respondent purported to cross-appeal on the grounds which the Motions Judge decided adversely to him, but which were not reflected in the order. By notice of cross-appeal, dated 7 June 2001, the respondent asked that the order of the Motions Judge be varied to include a declaration that the decision of the CSC contravenes his rights under paragraph 2(b) of the Canadian Charter of Rights and Freedoms (being Part I of the Constitution Act, 1982, Schedule B, Canada Act 1982, 1982, c. 11 (U.K.) [R.S.C., 1985, Appendix II, No. 44]) (the Charter), an order that the appellants supply him with a copy of the software and an award of his costs.
[5]The deficiency in the order in appeal was not identified until these reasons were being prepared. The Court therefore scheduled a further hearing before the same panel which heard the appeal, to receive oral submissions from counsel respecting the most expeditious and just way of correcting the deficiency, so that the Court could dispose of all issues argued in the appeal.
[6]After hearing from both counsel, and with the consent of counsel for the appellants, the Court suggested that the respondent should move before the Motions Judge, pursuant to paragraph 397(1)(a) of the Federal Court Rules, 1998, to reconsider the order on the ground that it does not accord with the reasons given, and that subsequent pleadings in the appeal should be amended accordingly.
[7]On 20 January 2003, the Motions Judge made an additional order nunc pro tunc, the operative portion of which reads in part:
The Previous Order is hereby amended to add the following provisions:
| Paragraph 3 The applicant's requests for relief found in paragraphs 1(a), 1(b)(i) and 1(c) of the applicant's Amended Originating Notice of Motion dated June 26, 2000 are hereby denied. |
| Paragraph 4 Success being divided, there will be no order as to costs. |
[8]The relief sought by the applicant in paragraphs 1(a), 1(b)(i) and 1(c) of his amended originating notice of motion reads as follows:
1. (a) a declaration that the decision of the Respondents Correctional Service of Canada and Commissioner of Corrections ("CSC") contravenes the Applicant's constitutional rights pursuant to section 2(b) of the Charter of Rights and Freedom, 1982;
(b) a review of the decision by the Respondents CSC to deny the Applicant access to:
(i) a copy of the Offender Intake Assessment software (current version in operation), which includes the: Custody Rating Scale (CRS), the CSIR, and the Community Risk/Needs Management Scale, among other features; and
. . .
(c) the Applicant's costs of this Application on a solicitor-client basis;
[9]Subsequently, counsel for the parties having complied with the order to our satisfaction, the Court concluded that the cross-appeal had been properly constituted.
[10]The appeal requires this Court to interpret for the first time subsection 4(3) of the Access to Information Act, R.S.C., 1985, c. A-1, as amended (the Act) and section 3 of the Access to Information Regulations, SOR/83-507 (the Regulations), to determine whether computer software is a "record" and whether the appellants must create and supply to the respondent records that do not exist. For the reasons that follow, I have concluded that the conjoint effect of the provisions of the relevant legislation in this appeal do not require the appellants to provide the respondent with any of the records he requested in his prayer for relief and that the learned Motions Judge erred in law in requiring them to provide the data and the code book mentioned in the order. I will deal with the issues in the cross-appeal later.
Factual Background
[11]The Research Branch of CSC conducts statistical research and analysis on various subjects in the field of corrections. In order to facilitate such research and analysis, the Research Branch is allowed access to several classified databases of other agencies, such as the Royal Canadian Mounted Police (RCMP), the Canadian Police Information Centre (CPIC) and the National Parole Board (NPB). These databases contain confidential and sensitive information relating to crimes, convictions and personal information of victims and offenders. Access to these databases is restricted to personnel with enhanced security clearance. Typically, CSC obtains such information online through electronic links to computer networks of the agencies mentioned above. This information may be compiled "as needed" into a temporary database that is used for statistical research and analysis and destroyed after the research project is complete. However, some projects rely on "live data" contained in current RCMP, CPIC and NPB databases that are accessed directly and, therefore, in respect of these projects, CSC does not create its own new database.
[12]The respondent is a criminologist. He conducts research with respect to and is a critic of the Canadian penal system, the actions of the Government of Canada in relation to it and the policies that underlie the system itself. In two separate letters, both dated 17 March 1997, he requested from the appellants certain information related to his research, namely data (data), a code book (code book) to interpret the Data and software (software) to use the data on a personal computer. This is how the Motions Judge described them [at paragraph 2]:
| a) the 1992-93 CSC release cohort currently being used to recalibrate the (GSIR) (General Statistical Indicator of Recidivism) with personal identifiers deleted (such as name of inmate/parolee, FPS number, or full date-of-birth, however year of birth will not violate privacy). |
| b) The Code Book used to define and identify/locate the variables in each case. |
| c) A Copy of the Offender Intake Assessment software (current version in operation), which includes the: Custody Rating Scale (CRS), the GSIR, the Community Risk/Needs Management Scale among other features. |
[13]I deal first with the respondent's requested data. The evidence is that they do not exist but could be recreated.
[14]Laurence Motiuk is the Director General of Research at CSC. In these proceedings, he swore an affidavit on 9 June 1998 and filed it to support the position of CSC. He was cross-examined at length on his affidavit. In cross-examination, he explained what is involved in recreating the data. I reproduce below part of what he said:
98. . . . in order to create or recreate the file which Mr. Yeager has requested would require considerable work to not only to [sic] remount the [backup data] tapes for the RCMP, the National Parole Board and the Correctional Service of Canada, define the parameters for the selection criteria of those variables, and also to recreate that file, would take considerable resources, and technical time and expertise in order to do so.
. . .
146. . . . It's with some technical expertise that it's available, and resources, in order to reconstruct what was done before. Basically, you're doing the research project all over again. You're setting the parameters for the sampling frame, you're setting the points, you have to reopen data files in other agencies and down load the information, reconstruct the you know, the algorhythms and the code that was written, in order to do that.
. . .
| 244. Q. Then you go on to say it would require a significant dedication of CSC computers and personnel over an extended period of time -- |
| A. Yes, to do his, the one that was requested by Mr. Yeager, the specific request. |
| 245. Q. What do you mean, "significant"? |
| A. Significant? Well, one would have to recreate the sampling frame of which the original study was done. They would have remount the tapes or the -- |
| 246. Q. How long? |
| A. How long? I would say at least 40 days full-time of somebody dedicated to that work, plus having the person with that expertise replaced who is already expended in terms --. [Emphasis added.] |
[15]Even if the data existed, to address privacy and security concerns, the appellants would need to purge the data of all personal identifiers. Mr. Motiuk explained the additional time and effort that would be required to remove personal identifiers:
251. . . . if it existed and did not have to be recreated it would take probably two days, roughly, time to make sure that there were no personal identifiers, that the cross tab . . . We'd have to run an analysis on every variable in the base and make sure that there was no way of cross-referencing to identify a particular case, and purge the records.
| 252. Q. When you run a variable, what is actually done? You press a key and it's doing it by itself or does someone actually sit there and -- |
| A. No. A statistical research officer would run a programme to look at the cross-referencing of a variety of variables. and look at the cell frequencies. and look at each one of the intersection of every variable to make sure that there was no way that this individual could be identified by an external user. And, you know, referenced back in terms of being able to fill in fact filled records, to know who this individual was. [Emphasis added.] |
[16]Mr. Motiuk acknowledged that the specific data which the respondent requested on 17 March 1997 was "slightly different". In his affidavit, he stated that the "live data" exists on the computer networks of the agencies mentioned above, but not in a separate database which the appellants now maintain. Although Mr. Motiuk did not describe at length how the production of this specific data would impact on CSC's operations, he indicated at paragraph 23 of his affidavit that the preparation of the data would "draw significant human and computer resources from current and ongoing research project" and "entail approximately two weeks of work with associated dedication of computers." It is unclear whether or not his estimate includes the time required to purge the data of personal identifiers.
[17]I turn next to the code book which forms the second part of the respondent's request. In cross-examination on his affidavit, Mr. Motiuk described a code book as a listing of definitions of the variables contained in the database. Although code books are occasionally produced by researchers at CSC, none was produced in relation to the data which the respondent had requested. According to Mr. Motiuk, code books are often unnecessary because CSC researchers are well acquainted with the variables, definitions and underlying data structures. By omitting this "very labour intensive stage" of a project, researchers can save themselves "months of work". He also stated that the creation of a code book to accompany the data would require "a good week to two weeks of text entry into a computer".
[18]The respondent's final request was for the software. With respect to this request, Mr. Motiuk deposed that none of the software requested by the respondent exists in a format for use on a personal computer. Rather, it exists in a format for use on a mainframe computer system. In cross-examination, he explained in detail that all the software which the respondent had requested was either obsolete and, therefore, discarded; inseparable into independent modules; or subject to copyright licensing agreements.
[19]The Motions Judge described the software as a massive repository of text files that includes many different integrated components or modules, the development of which was a multi-million dollar project that involved a number of outside software firms.
[20]On 10 April 1997 the appellants refused the respondent's request for the data, code book and the software, claiming that the Act does not impose an obligation to create records that do not exist. The respondent then complained unsuccessfully to the Information Commissioner respecting the answers he had received from CSC. Pursuant to section 41 of the Act the respondent applied to the Trial Division for a judicial review of the appellants' decision denying his request.
Relevant Legislative Provisions
[21]It might be useful to set out here the relevant legislative provisions in the Act [ss. 2(1), 3, 4(1)(a), (b), (3), 12(1), 41, 53(1), (2)]. They read:
2. (1) The purpose of this Act is to extend the present laws of Canada to provide a right of access to information in records under the control of a government institution in accordance with the principles that government information should be available to the public, that necessary exceptions to the right of access should be limited and specific and that decisions on the disclosure of government information should be reviewed independently of government.
. . .
3. In this Act,
. . .
"record" includes any correspondence, memorandum, book, plan, map, drawing, diagram, pictorial or graphic work, photograph, film, microform, sound recording, videotape, machine readable record, and any other documentary material, regardless of physical form or characteristics, and any copy thereof;
. . .
4. (1) Subject to this Act, but notwithstanding any other Act of Parliament, every person who is
(a) a Canadian citizen, or
(b) a permanent resident within the meaning of the Immigration Act,
has a right to and shall, on request, be given access to any record under the control of a government institution.
. . .
(3) For the purposes of this Act, any record requested under this Act that does not exist but can, subject to such limitations as may be prescribed by regulation, be produced from a machine readable record under the control of a government institution using computer hardware and software and technical expertise normally used by the government institution shall be deemed to be a record under the control of the government institution.
. . .
12. (1) A person who is given access to a record or a part thereof under this Act shall, subject to the regulations, be given an opportunity to examine the record or part thereof or be given a copy thereof
. . .
41. Any person who has been refused access to a record requested under this Act or a part thereof may, if a complaint has been made to the Information Commissioner in respect of the refusal, apply to the Court for a review of the matter within forty-five days after the time the results of an investigation of the complaint by the Information Commissioner are reported to the complainant under subsection 37(2) or within such further time as the Court may, either before or after the expiration of those forty-five days, fix or allow.
. . .
53. (1) Subject to subsection (2), the costs of and incidental to all proceedings in the Court under this Act shall be in the discretion of the Court and shall follow the event unless the Court orders otherwise.
(2) Where the Court is of the opinion that an application for review under section 41 or 42 has raised an important new principle in relation to this Act, the Court shall order that costs be awarded to the applicant even if the applicant has not been successful in the result. [Emphasis added.]
[22]Section 3 of the Regulations reads:
3. For the purpose of subsection 4(3) of the Act, a record that does not exist but can be produced from a machine readable record under the control of a government institution need not be produced where the production thereof would unreasonably interfere with the operations of the institution. [Emphasis added.]
Reasons of the Motions Judge
[23]The Motions Judge concluded that the data and the code book were records that should be produced because to do so would not unreasonably interfere with the appellants' operations. After noting that the appellants had conceded that the data can be recreated, she concluded that the appellants bore the onus of establishing that producing the data would unreasonably interfere with operations and that they failed to meet that burden. She acknowledged that Mr. Motiuk believed the production of the data would require "significant . . . resources" and "approximately two weeks worth of work", but stated (at paragraph 12) that:
No evidence was put forward about the operations of CSC or its Research Branch. I therefore have no idea whether it has a large staff and a complex computer system with many links to the systems operated by the RCMP and the National Parole Board or whether Mr. Motiuk works with a small staff and limited computer capacity. Further, there was no evidence about how many staff members and how much time would be required to do the two weeks' work he thinks will be needed. As well, no information was provided about the extent to which CSC's computers would be occupied creating the Requested Data. I also was given no indication whether any important or urgent matters are presently being researched by the Research Branch and what impact, if any, the creation of the Requested Data would have on the workload of the Research Branch. [Emphasis added.]
[24]The Motions Judge also concluded that a code book was capable of being produced, using computers, software and expertise normally used by the appellants. Although Mr. Motiuk stated that a code book would take "a good two weeks" to produce, she felt that the appellants had adduced no evidence to satisfy her that this work would unreasonably interfere with their operations. She decided, therefore, that the respondent was entitled to access to both the data and a code book.
[25]She accepted that the software does not exist in a format for use on a personal computer. As a result she concluded that it was not producible because writing software is not normally done by the appellants but, rather, by outside consultants. Furthermore, she was of the opinion that the software requested was not a "record" as defined by the Act. Consequently, she decided that the respondent had no right of access to it.
Issues
[26]The appeal raises two issues:
(1) whether the appellants were obliged, pursuant to subsection 4(3) of the Act, to create and provide to the respondent the requested data and a code book of technical terms, simply because they were capable of recreating such records; and
(2) whether the learned Motions Judge erred in concluding that the appellants had not met the onus imposed by section 3 of the Regulations of showing that the production of these records would unreasonably interfere with the appellants' operations.
Analysis
The Interpretive Framework
[27]I begin the analysis by reproducing the following well-known passage from Ruth Sullivan, Driedger on the Construction of Statutes, 3rd ed. (Toronto: Butterworths, 1994), at page 131:
There is only one rule in modern interpretation, namely, courts are obliged to determine the meaning of legislation in its total context, having regard to the purpose of the legislation, the consequences of proposed interpretations, the presumptions and special rules of interpretation, as well as admissible external aids. In other words, the courts must consider and take into account all relevant and admissible indicators of legislative meaning.
[28]On this interpretive approach, one of the primary considerations is, of course, the purpose of the Act. Section 2 of the Act states clearly that its purpose is to "provide a right of access to information in records under the control of a government institution". In Dagg v. Canada (Minister of Finance), [1997] 2 S.C.R. 403, at paragraph 61, La Forest J. explained its purpose in the following passage of his reasons:
The overarching purpose of access to information legislation, then, is to facilitate democracy. It does so in two related ways. It helps to ensure first, that citizens have the information required to participate meaningfully in the democratic process, and secondly, that politicians and bureaucrats remain accountable to the citizenry.
In that context, the Act mentions specifically three principles that should guide this Court's interpretation of its provisions: (1) that government information should be available to the public; (2) that necessary exceptions to the right of access should be limited and specific; and (3) that decisions on the disclosure of government information should be reviewed independently of government.
[29]There is no dispute that the respondent is either a Canadian citizen or permanent resident of Canada or that CSC is a government institution. Therefore, subsection 4(1) of the Act states unambiguously that the respondent "has a right to and shall, on request, be given access to any record under the control of" the appellant. The issue to be resolved, therefore, is whether each of the data, a code book and the software is a "record" under CSC's control.
The Definition of a "Record"
[30]Section 3 of the Act defines a "record" as including:
3. . . . any correspondence, memorandum, book, plan, map, drawing, diagram, pictorial or graphic work, photograph, film, microform, sound recording, videotape, machine readable record, and any other documentary material, regardless of physical form or characteristics, and any copy thereof.
[31]By this definition, the data and a code book requested by the respondent would each be a "record" for the purposes of the Act, but neither exists. In order to satisfy the respondent's request, the data would have to be recreated and modified, and a code book would have to be prepared. Although the software does exist, it is less clear whether the software is a "record".
The Obligation to Produce a Non-Existent Record
[32]For convenience, I reproduce again subsection 4(3) of the Act. It reads:
4. . . .
(3) For the purposes of this Act, any record requested under this Act that does not exist but can, subject to such limitations as may be prescribed by regulation, be produced from a machine readable record under the control of a government institution using computer hardware and software and technical expertise normally used by the government institution shall be deemed to be a record under the control of the government institution. [Emphasis added.]
[33]In my opinion that subsection provides that a non-existent record that can be produced from an existing machine readable record is deemed to be a record to which the respondent is entitled access. Not surprisingly, the parties have markedly differing opinions respecting the obligations imposed by subsection 4(3).
[34]The appellants claim that the Motions Judge erred in law by determining that subsection 4(3) requires them to create items that do not exist in any form. They submit that subsection 4(3) is intended to apply to situations where a "record" exists only in machine readable form, and in that case, requires the record to be produced in a more accessible form. For example, the appellants would accept that they are required to print an existing machine readable record and provide the paper copy to the respondent. They insist, however, that there is no obligation to create a record unless that record already exists in machine readable form. According to the appellants, there is no obligation to adapt, manipulate, change or alter the existing machine readable record. The appellants submit that any other interpretation would be contrary to the purpose of the Act, which is to grant access to records held by government institutions but not to conjure up records solely for the purpose of satisfying requests.
[35]For their part, the respondent says that this interpretation is too restrictive and not in keeping with the purpose of the Act as explained in the jurisprudence.
[36]I have concluded that the interpretation of subsection 4(3), for which the appellants contend, is not plausible. The interpretation proposed by the appellants is in stark contrast with the express language of this provision, contrary to accepted principles of statutory interpretation and unsupported in light of the overarching purpose of access to information legislation as stated in section 2 of the Act and explained by the Supreme Court of Canada.
[37]First, the appellants' suggestion that subsection 4(3) applies to a record that already exists in machine readable form is implausible. If a record exists in machine readable form, then it cannot be a record that does not exist. Such an interpretation would, therefore, be completely illogical. Second, the appellants' interpretation is contrary to the express language of the Act. Subsection 4(3) applies where a record "does not exist". The subsection does not qualify this reference with the clause "does not exist other than in machine readable form" or words of similar import. Third, the appellants' interpretation would render subsection 4(3) devoid of any meaning or effect. Since a "record" is expressly defined in section 3 of the Act to include a machine readable record, the respondent has a right to access such a record regardless of subsection 4(3). That is, subsection 4(1) already contemplates that a machine readable record must be supplied to the respondent.
[38]Thus, the appellants' interpretation is contrary to accepted principles of statutory interpretation, including the "plausible meaning rule", the "presumption of linguistic competence" and the "rule against tautology" (see R. Sullivan, supra, at pages 103, 157 and 159). The only plausible interpretation of subsection 4(3) requires the appellants to produce a record that "does not exist" in any form, machine readable or otherwise. In enacting subsection 4(3) Parliament must have contemplated two different records: a new and distinct record must be produced from an existing machine readable record.
[39]Moreover, the interpretation preferred by the respondent is more consistent with the "overarching purpose" of the Act, which is to "facilitate democracy" by ensuring participation by citizens and accountability of politicians and bureaucrats (see Dagg, supra). In fact, the respondent seeks access to information for precisely the reasons contemplated by the Supreme Court of Canada, namely, to participate in and ensure the accountability of the Canadian government's research and to analyse the Canadian penal system.
[40]Having concluded that subsection 4(3) contemplates the production of a new record, it is still necessary to determine the meaning of the phrase "from a machine readable record". In my opinion, the answer to this question is largely contextual and fact-specific. Whether a record is indeed producible "from" a machine readable record depends upon a number of factors, including the requisite amount of independent composition as compared to purely mechanical and routine editing or manipulation. But it is necessary here to state the limitations on the obligation of government institutions to produce otherwise non-existent records, as discussed below.
Limitations on the Obligation to Produce a Non-Existent Record
[41]First, subsection 4(3) itself states that non-existent records must be produced only where they are capable of being produced "using computer hardware and software and technical expertise normally used by the government institution". It is not questioned that the CSC has the ability to recreate the data and the code book.
[42]Second, the obligation to produce non-existent records is "subject to such limitations as may be prescribed by regulation". Limitations are prescribed by section 3 of the Regulations, which provides that:
3. . . . a record that does not exist but can be produced from a machine readable record under the control of a government institution need not be produced where the production thereof would unreasonably interfere with the operations of the institution. [Emphasis added.]
[43]The Motions Judge was of the opinion that the appellants adduced no evidence to demonstrate that satisfying the respondent's request would unreasonably interfere with its operations. For this reason she ordered the appellants to supply to the respondent a copy of the data and a code book.
[44]This is how she explained it in paragraph 12 of her reasons:
No evidence was put forward about the operations of CSC or its Research Branch. I therefore have no idea whether it has a large staff and a complex computer system with many links to the systems operated by the RCMP and the National Parole Board or whether Mr. Motiuk works with a small staff and limited computer capacity. Further, there was no evidence about how many staff members and how much time would be required to do the two weeks' work he thinks will be needed. As well, no information was provided about the extent to which CSC's computers would be occupied creating the Requested Data. I also was given no indication whether any important or urgent matters are presently being researched by the Research Branch and what impact, if any, the creation of the Requested Data would have on the workload of the Research Branch. [Emphasis added.]
[45]Where evidence on an issue exists and should be evaluated for sufficiency, but the Motions Judge concludes that there is "no evidence" on the issue, then the conclusion is one of law. Housen v. Nikolaisen, [2002] 2 S.C.R. 235 teaches that such a conclusion should be reviewed on a standard of correctness.
[46]In determining whether satisfying the respondent's requests would "unreasonably interfere" with the appellants' operations, as that term is used in section 3 of the Regulations, it is my respectful view that the Motions Judge ignored significant evidence in the record. In doing so she was clearly wrong to conclude that there was no evidence in the record to mandate a finding that the production of the non-existent records sought by the respondent would unreasonably interfere with the appellants' operations. On any objective view of the whole of the evidence, I am of the opinion that production of the particular items requested by the respondent would constitute unreasonable interference with the operation of the appellants.
The Data
[47]Before the Motions Judge the appellants conceded that the data requested by the respondent can be produced in the manner contemplated by subsection 4(3). That is, the requested record (data without personal identifiers) can be produced from a machine readable record (data with personal identifiers) using computer hardware, software and technical expertise normally used by CSC.
[48]However, there was abundant evidence in the record to conclude that the production of the otherwise non-existent data would unreasonably interfere with the appellants' operations. The Motions Judge inaccurately stated that "[n]o evidence was put forward about the operations of CSC or its Research Branch." At paragraph 7 of her reasons, she in fact quoted extensively from portions of Mr. Motiuk's affidavit describing CSC, the whole of which clearly explains the role of the Research Branch and its computer links, software and databases.
[49]The Motions Judge suggested that she had [at paragraph 12] "no idea whether it has a large staff and a complex computer system . . . [or] a small staff and limited computer capacity." However, on cross-examination, Mr. Motiuk expressly stated in response to question 356 that "[t]here are many databases in the Correctional Service of Canada. There's 1200 employees and many researchers." This is merely one of several pointed statements by Mr. Motiuk on cross-examination. A thorough review of the record indicates that there is, in fact, an abundance of information about CSC's computer and human resources.
[50]She also stated that "there was no evidence about how many staff members and how much time would be required to do the two weeks' work he thinks will be needed." With respect, she took Mr. Motiuk's comment in complete isolation from the rest of his affidavit and his testimony on cross-examination. Mr. Motiuk's remark must be viewed in light of his testimony as a whole. Viewed in this way, it is clearly unreasonable to expect anyone to undertake the work that Mr. Motiuk suggested as being necessary to recreate the record that would not otherwise exist.
[51]She also inaccurately commented that [at paragraph 12] she "was given no indication whether any important or urgent matters are presently being researched". In my view, however, given the detail in which Mr. Motiuk explained CSC's role, it is evident that the CSC is currently engaged in a host of ongoing projects. Contrary to her finding, it is easy to see "what impact, if any, the creation of the Requested Data would have on the workload of the Research Branch." As Mr. Motiuk stated on cross-examination in response to questions 381-383:
. . . we do not have the resources because there is costs involved, staff time, and resources and equipment that are required in order to create entities such as that, files such as that, for external use. . . . We are not funded to--we have no discretionary funds to give to outside researchers. We're an applied operational unit.
[52]Finally, when one looks at the record as a whole it is clear that fulfilling the respondent's request would unreasonably interfere with the appellants' operations. In cross-examination, Mr. Motiuk explained at length CSC's role, activities, procedures and policies generally (at questions 6-9, 12-18, 77-78, 97-100, 103-184, 226-287, 305-314, 344-384, 389-414, 421-428) and in particular with respect to the work involved in recreating data used for research projects (at questions 97-100, 143-144, 157-164, 231, 244-254, 258-262, 344-351, 359-361, 381-382). Indeed, most of his affidavit and testimony on cross-examination is directed to this issue.
[53]Mr. Motiuk explained broadly the effort necessary to recreate records such as those sought by the respondent. In other places, he described specifically the implications of fulfilling the respondent's particular requests. The Motions Judge, however, was extremely selective in the evidence she considered, referring at paragraph 10 of her analysis only to a brief excerpt from Mr. Motiuk's affidavit concerning the effort required to fulfill the respondent's request. While additional contextual evidence might have been helpful to the Motions Judge, her task was to examine all the evidence presented to determine whether the appellants had met the statutory burden. Instead, she was selective in the evidence she considered and thus overlooked other evidence which was clearly relevant. In doing so she was clearly wrong.
[54]Furthermore, she placed too high a burden upon the appellants to justify their refusal. She failed to give sufficient weight to all relevant considerations, notably the preponderance of evidence contained in the record that supports the appellant's position. To my mind it is inconsistent with the principle that "[r]ights to state-held information are designed to improve the workings of government; to make it more effective, responsive and accountable" (see Dagg, supra, at paragraph 63) to require a government institution to adduce more detailed and precise contextual evidence of its operations than the appellants did in this case. Where a motions judge has failed to give sufficient weight to all relevant considerations, an appellate court is entitled to intervene, Reza v. Canada, [1994] 2 S.C.R. 394.
A Code Book
[55]The appellants claim that the code book does not exist in any form, mechanical or otherwise, and, therefore, need not be produced. In light of my interpretation of subsection 4(3), I do not agree with the appellants on this point. Since a code book can be produced from either the underlying or requested data, which are clearly machine readable records, using the computer hardware and software and technical expertise normally used by the appellants, a code book should be produced unless its production unreasonably interferes with the appellants' operations.
[56]The appellants attempt to distinguish between the data and a code book. The appellants conceded that the data was producible because it already exists in machine readable form; it did not require the creation of an entirely new record. In fact, however, the data as requested by the respondent does not already exist in machine readable form. Before it can be produced as requested by the respondent, it would need to be recreated and purged of personal identifiers. Thus, the data as requested by the respondent is distinct from the existing machine readable record. The difference between the data as requested by the respondent and a code book is merely one of degree. Production of the data would require only manipulation of the distinct underlying machine readable record; while production of a code book would require a somewhat greater investment of time and effort, but would nevertheless be "from" the underlying machine readable record. Consequently, the appellants have not satisfied me that the Motions Judge erred in this respect.
[57]The fact that a greater investment of time and effort is required to create a code book than the data does not, by itself, relieve the appellants of their obligation under subsection 4(3) to produce such a record. However, the greater amount time and effort required to fulfill the respondent's request for a code book does directly impact upon a determina-tion of the unreasonableness of the interference caused to the appellants' operations. In considering whether the production of a code book would unreasonably interfere with the appellants' operations, the Motions Judge again cited a lack of evidence adduced by the appellants. For reasons similar to those that I have discussed, I have concluded that there was clearly sufficient evidence in the record to determine that producing a code book would unreasonably interfere with the appellant's operations.
[58]In response to questions 288-304 on cross-examination of his affidavit, Mr. Motiuk detailed the work involved and the consequences of fulfilling the respondent's request for a code book. As I have already mentioned, he explained that code books are often unnecessary to CSC research, and by omitting this "very labour intensive stage" of a project, researchers can save themselves "months of work". In particular, he stated that producing a code book to accompany the particular data requested by the respondent would require "a good week to two weeks of text entry into a computer of rewriting variable labels and value labels in terms of the definitions".
[59]For these reasons I am of the opinion that the appellants must succeed.
[60]I pass now to consider the issues which form the subject-matter of the cross-appeal.
The Software
[61]On 17 March 1997 the respondent made a request for a copy of the software. The Motions Judge decided that the software was not producible using computer hardware and software and technical expertise normally used by the appellants. She held that since it existed only for use on a mainframe computer, but not a personal computer, the appellants had no obligation to produce the latter format. That task would normally be performed by external agencies. The respondent cross-appeals this finding, arguing that the Motions Judge erred by assuming that he necessarily required the software in a format for use on a personal computer.
[62]In my view, the contention of the respondent on this point is not without merit. The evidence of the respondent is not that he required the software for use on a personal computer. Indeed, the respondent claims he had access to a mainframe computer. The primary question, therefore, is not whether the appellants have an obligation under subsection 4(3) to produce the software in a format for use on a personal computer (although, clearly, such work is not normally done by the appellants), but whether the mainframe computer software is an existing "record" as defined under section 3 of the Act.
[63]The Motions Judge held that the software was not a "record". On this point, I am in respectful agreement with her interpretation of section 3 of the Act. She concluded that the software is an item used to generate, view or edit a record, as opposed to a record itself. The software is not analogous to any of the items listed in the definition under section 3. In her attempt to interpret Parliament's intention, she explained that if computer software was contemplated in the definition of a "record" it would have been expressly mentioned. In light of the language of section 3 and the overall purpose of the Act, her approach is consistent with accepted principles of statutory interpretation, and I can find no error in her conclusion. I would only add that, the software is not only not a "record", it is not "under the control of a government institution". It is under the control of the external developer, and the appellants are mere licensees. They have no authority to copy or otherwise use the software outside of their agreements.
Infringement of paragraph 2(b) Charter Rights
[64]The respondent submits that the Motions Judge erred in refusing to declare that the purpose and effect of the appellants' actions were to deny his constitutional right to freedom of expression as guaranteed by paragraph 2(b) of the Charter. The respondent claims that he did not seek a declaration that the Charter provides him with a right of access to information, but merely that the appellants infringed his constitutional right.
[65]I have not a little difficulty understanding the distinction the respondent attempts to draw. Unless the respondent has a constitutional right of access to information, the appellants could not have violated that right. The Motions Judge considered the decision of the Ontario Divisional Court in Ontario (Attorney General) v. Fineberg (1994), 19 O.R. (3d) 197, which held that paragraph 2(b) does not provide a general right of access to information. Without endorsing all the reasons for decision given in that case, I am in respectful agreement with the conclusion of the Motions Judge that the respondent's Charter right was not contravened here.
Costs
[66]The respondent raises the issue of costs pursuant to section 53 of the Act, which I again reproduce for convenience here:
53. (1) Subject to subsection (2), the costs of and incidental to all proceedings in the Court under this Act shall be in the discretion of the Court and shall follow the event unless the Court orders otherwise.
(2) Where the Court is of the opinion that an application for review under section 41 or 42 has raised an important new principle in relation to this Act, the Court shall order that costs be awarded to the applicant even if the applicant has not been successful in the result.
[67]The Motions Judge stated [at paragraph 34] that "[a]s success was evenly divided, there will be no order as to costs." Clearly, subsection 53(1) of the Act gives the Motions Judge discretion in awarding costs. However, subsection 53(2) qualifies that discretion. I can only infer that, since costs were not awarded to the applicant (the respondent in this proceeding), the Motions Judge was of the opinion that the respondent did not raise an important new principle in relation to the Act.
[68]As stated at the outset of these reasons and accepted by both parties, the appeal and cross-appeal require this Court to consider subsection 4(3) of the Act and section 3 of the Regulations for the first time. The respondent contends that the issues raised in the appeal and cross-appeal involve an important new principle in relation to the Act. He contends further that the implications of these issues are enormous not only for government institutions but also for the public. In response, the appellants contend that simply because an issue is raised for the first time does not make it an important new principle. This may be true, but to my mind, if the issues raised by the respondent in this case do not constitute important new principles, it is difficult to imagine one that would. If the Motions Judge thought otherwise, then I am of the opinion that she was wrong in law. I agree with the respondent respecting the implications of this case. It does involve an important new principle. Consequently although the respondent was ultimately unsuccessful in this appeal, and in his cross-appeal, nevertheless, in my opinion, he should be awarded his costs throughout.
Disposition
[69]I would allow the appeal, dismiss the cross-appeal, and dismiss the application for judicial review. Pursuant to subsection 53(2) of the Act the respondent should have his costs, both here and below, on a party-and-party basis.
Stone J.A.: I agree.
* * *
The following are the reasons for judgment rendered in English by
[70]Malone J.A. (dissenting in part): I dissent from the majority reasons only in connection with paragraph 69 and the issue of costs. Subsection 53(2) of the Act reads as follows:
53. . . .
(2) Where the Court is of the opinion that an application for review under section 41 or 42 has raised an important new principle in relation to this Act, the Court shall order that costs be awarded to the applicant even if the applicant has not been successful in the result.
[71]Once it has been determined, as in this application, that an important new principle in relation to the Act has been raised by the respondent, his lack of success in the initial application and on his cross-appeal becomes irrelevant. Costs cannot be awarded against him and the only remaining issue is the question of costs to be awarded in his favour.
[72]In my view, a lump sum award under subsection 400(4) of the Federal Court Rules, 1998 best reflects the true intention of Parliament in enacting this unusual subsection, namely, that the respondent should be awarded his reasonable costs and not the normal amount generated by an award of party-and-party costs under the Tariff.
[73]I would set the respondent lump sum costs at $20,000 inclusive of GST and all disbursements.