Access to Information Orders
Decision Information
• Orders of the Landlord and Tenant Board and the Ontario Rental Housing Tribunal containing identified terms.
• Section 57(1) (fee estimate) - partly upheld.
• Fee estimate upheld only to the extent of $344.00 in record preparation time and photocopy charges of $137.60.
• Institution to issue final access decision to the appellant.
Decision Content
NATURE OF THE APPEAL:
The Landlord and Tenant Board (the Board) received a request under the Freedom of Information and Protection of Privacy Act (the Act) from an individual seeking orders or decisions that are located in the computer records of the Board and its predecessor, the Ontario Rental Housing Tribunal (the ORHT). The requester seeks access to orders or decisions that contain certain identified keywords or terms set out in his request, for the time period from January 1, 1988 to April 18, 2008.
In its initial decision letter, the Board advised the requester that its case management computer system could not be searched by keywords. However, the Board provided the requester with three of its decisions that had been located by one of its solicitors dealing with one of the terms specified by the appellant in his request.
The requester (now the appellant) appealed the Board’s decision. In his Notice of Appeal, the appellant took the position that the Board was obliged to provide an explanation for why its computer system is not searchable. The appellant stated that this would include the name and contact information for the company or individual that maintains the Board’s computers and database. The appellant also asserted that the Board should provide information about alternative methods for searching its case law and that if a software update was required to allow a keyword or term search this should be done immediately at the Board’s expense. The appellant also stated that, if it was not possible to conduct a computerized search of its decisions, a manual search should be conducted, or alternatively, the Board should allow the appellant to conduct his own review.
At the intake stage of the appeals process, the Board advised the Information and Privacy Commissioner (the IPC) that it did not have a mechanism to conduct an electronic search for responsive records based on keywords and that a manual search would be too time-consuming. The Board also indicated that, because the decisions contained the personal information of individuals other than the appellant, it would not be feasible for the appellant to personally review the records in their database.
At mediation, the appellant was advised of the Board’s position. The appellant was not satisfied with the Board’s explanation. At the Mediator’s request the Board then provided a supplementary decision letter relating to a manual search for responsive records, setting out that:
- it did not have access to Court decisions issued under the Landlord and Tenant Act, and, therefore, could only search for orders of the ORHT and the Board after June 16, 1998, the day that the ORHT first became operational.
- the only way to search for orders based on keywords was to search its case management system by application type for files where one would expect the orders to contain the particular keywords, and to conduct a manual review of all orders in each relevant application type.
- the estimated fee for conducting this type of search and for preparing responsive records for disclosure was the sum of $86,611.60.
- it required a deposit of 50% of the estimated fee before proceeding with the request.
At mediation, the appellant advised the Mediator that he is challenging the fee estimate. The Board also indicated that it would be claiming the mandatory exemption at section 21(1) of the Act (invasion of privacy) for any personal information in the records. This was, in effect, an “interim access decision” (see Orders P-81, PO-2634), and the proposed exemption claim is not at issue in this appeal, whose focus is the appellant’s objection to the fee estimate.
Mediation did not resolve the appeal and it was moved to the adjudication stage of the process, where an adjudicator conducts an inquiry under the Act.
After the initial exchange of representations in the inquiry, the Board forwarded correspondence to this office advising that it had recently adopted a search engine that allowed for some of its decisions to be searched by keyword or term. This generated further representations which are addressed in more detail below.
The time frame of the appellant’s request includes court decisions issued under the Landlord and Tenant Act that predate the creation of the OHRT. As already noted, the Board does not have this body of case law in its possession. For the purposes of the discussion that follows, I will be considering those records that are in the Board’s custody or under its control, which consist of decisions issued by the ORHT since its inception in June 1998 until it was replaced by the Board, and subsequent decisions issued by the Board up to the date of the appellant’s request. This appeal does not address decisions issued by the courts before the inception of the ORHT.
BACKGROUND
At its inception on June 17, 1998, the ORHT began using Evans Caseload computer software for its file management system. The Board submits that at the time, Evans Caseload offered the most functionality for the price, but it did not provide the ability to search for decisions by keywords. The Board initially took the position that the only way to search its orders based on specific terms or keywords was to search the Evans Caseload case management system by application type for application files where one would expect to find orders that contain the required terms. Then each order found within those application files would have to be read to identify whether they actually contained those terms. The estimated total fee of $86,611.60 was made up of search time in the sum of $86,130.00, record preparation time in the sum of $344.00 and copying charges in the sum of $137.60.
As the inquiry in this appeal came to a close, the Board changed its position about searching its decisions. The Board advised that it had recently introduced a new feature on its website which makes “redacted” versions of some contested Board orders available to the public. The Board advised, however, that it would take some time before the full range of contested orders issued under the Residential Tenancies Act (RTA) would be available through the use of this feature.
SUMMARY OF CONCLUSIONS
The essential issue in this appeal is whether the Board’s fee estimate should be upheld, in whole or in part. The Board claims that this fee represents the cost of fulfilling the appellant’s request.
For his part, the appellant argues that it is a simple matter to obtain free or inexpensive computer software to accomplish the necessary search, without unreasonably interfering with the Board’s operations. He also takes the position that the Board’s recent process of creating a searchable database, which would add orders over time and would not include the orders issued under the Tenant Protection Act, does not adequately respond to his request.
The Board did not attempt a simple database search along the lines suggested by the appellant, and could only suggest expensive and complex solutions to respond to the appellant’s request.
I conclude that by failing to apply a simple, reasonable and cost effective solution that is readily available, the Board has failed to establish its entitlement to most of its fee estimate.
DISCUSSION
The Representations
The Board’s Representations
The Board initially submitted that because an order is produced using the Board’s case management system the resulting order is part of the electronic application file. The Board stated that, accordingly, it is only possible to search for responsive decisions manually.
The Board submits that its Freedom of Information Coordinator (the Coordinator) formulated her fee estimate for a manual search by considering the application types most likely to contain the requested keywords and projecting the number of decisions that would have to be reviewed to identify those responsive records that contained the keywords. She then used workload reports for the Board and the ORHT to determine the number of decisions that related to the relevant application types. The Board submits that the Coordinator adjusted the overall numbers in the worksheets based on representative samples, where appropriate. In situations where the worksheets could not provide the required information, the Board states that the Coordinator used searches on its computer system to estimate the number of decisions that would have to be reviewed. In support of its position on the fee estimate, the Board referred the supplementary decision sent to the appellant that contained the fee estimate.
As set out in the supplementary decision letter, considering the types of applications before the ORHT and the Board, the Coordinator estimated that 172,259 orders would have to be reviewed for the keywords sought by the appellant. In an effort to estimate the time required to review all of those orders to verify that they contained the keywords sought by the appellant, based on a representative sample of potentially relevant application types, the Coordinator timed how long it took to read the orders. She estimated that she could review 60 orders per hour. In her view, it would therefore take 2781 hours to review the 172,259 orders. Based on $30 per hour of search time, she estimated a fee of $86,130 for search time. She also estimated $344.00 in record preparation time as well as photocopy charges of $137.60. Accordingly, the total estimated fee was the sum of $86,611.60.
The Board submits that developing a computer system to retrieve decisions that include specific keywords would require storage equipment and technical expertise not normally used by the Board and would unreasonably interfere with its operations.
In an effort to establish that a computer search was not possible, the Board relied on a memorandum provided by its Applications Services Unit Manager that accompanied its representations. In the memorandum, the Board’s Applications Services Unit Manager states that there are two common methods of achieving searchable documents by keywords:
- Database document imaging (by storing documents within a database table in a binary format) and,
- Metadata document imaging (by using third party document management software).
He states that both of these methods are outside the normal scope of expertise of the Board’s Information Technology (IT) staff.
In particular, he sets out the following in the memorandum:
Database document imaging:
At present, the [Board’s] database structure does not allow decision documents (Orders) to be searchable via keywords. The Orders are NOT stored within a relational database; instead the database simply stores a pointer to the location of the order on the file server. This is the technology that we use that is provided through a 3rd party software vendor, Evans Caseload.
Metadata document imaging:
An alternative to a database search is to search documents stored on a file server with a third-party document management solution. Currently, [the Board] does not have this solution in place, therefore this solution will require time, resources and IT infrastructure to implement.
It is estimated that such a solution would require the [the Board] to acquire software licenses and the services of a 3rd party provider at a cost of approximately $75,000.00 to $100,000.00 and an elapsed time of 6-9 months following OPS governance re: obtaining software licenses and hiring a 3rd party vendor through the RFS process. As well, it would require the use of senior IT expertise from the Community Services I & IT Cluster (CSC), at a time when these resources are devoted to high-priority Government IT initiatives.
Conclusion from IT staff:
Storing of documents in a database table is very time consuming due to the overhead created on the database itself. Storing and searching for records in this method is very time consuming and creates a situation that puts excessive strain on computer and network resources. Therefore the database search solution is not recommended for this activity.
Metadata document imaging is the more modern approach to search stored documents on a file server. However, as indicated above, to achieve this, [Board] management would require resources and expertise not currently available to the [Board] and implementing this solution would have a detrimental impact on current high-priority [Board] IT projects. [Emphasis in original]
The Board submits that it (and the Community Services I & IT Cluster) would face similar technical and resource issues to those described in Order P-1572 if it were to attempt to develop a keyword search system.
In Order P-1572, former Assistant Commissioner Tom Mitchinson addressed a request for, amongst other things, bulk access to the data elements of a database. He upheld the Ministry of Consumer and Commercial Relations’ decision refusing access, writing:
Having reviewed the representations of the Ministry, along with the supporting affidavits, I am persuaded that production of the record would require the use of computer hardware, software and technical expertise not normally used by the Ministry in the operation of its programs. I also accept the Ministry’s evidence that, even if the record could be produced from the system normally used by the Ministry, to do so would unreasonably interfere with the operations of the Ministry. Therefore, I find that, in accordance with section 2 of O. Reg. 460, the ONBIS data elements in bulk form do not satisfy the definition of a “record” under section 2 of the Act. Accordingly, the requested record is not accessible under the Act.
Former Assistant Commissioner Mitchinson provided a lengthy postscript to the order which commented on the Ministry of Consumer and Commercial Relations’ failure to turn its mind to access-related implications when it developed the ONBIS system:
Historically, when business registration filings were made in paper form and recorded by the Ministry on microfilm or microfiche, information similar to that at issue in this appeal was available to the public in bulk format. At that time, these microfiche and microfilm records could be reproduced by the Ministry with relative ease and were made available to purchasers at low cost. (See my Order P-1114). There is more than a little irony in the fact that the efficiencies of electronic filing, storage and retrieval have become a barrier to access to business registration information in bulk format in this particular case.
I have found that producing the requested record in this case would require resources not normally used by the Ministry and would unreasonably interfere with the Ministry’s operations. However, I also note that the total software development, hardware acquisition and staffing costs to produce the record amount to approximately one percent of the total cost of developing the ONBIS system over a period of five years, based on the Ministry’s own estimates. While I have no specific evidence on this, I believe it is reasonable to assume that these costs would have been considerably lower if the Ministry had originally designed and developed software components in the ONBIS database which would permit the appropriate data extraction for emulation of bulk records previously available to the public under the Act or otherwise.
Although it is arguably understandable that the Ministry did not turn its mind to these access-related implications when it developed the ONBIS system several years ago, the results of this order clearly demonstrate how this lack of foresight can impact on public access to records. As the Ministry and other parts of government become increasingly reliant on electronic databases such as ONBIS to deliver their programs, it is critically important that public accessibility considerations be part of the decision-making process on any new systems design.
This issue is not unique to Ontario. It has been raised in the past by former federal Information Commissioner John Grace and others, and is a serious concern of access to information professionals in all jurisdictions.
The public’s statutory right of access to government records is a critically important component of our system of government accountability. Accessibility and transparency are inexorably linked to public trust and faith in government. Retaining access rights to raw electronic data is an important part of this overall accountability system, and factoring public access requirements into the design of new systems will ensure that these important rights are in fact enhanced rather than irretrievably lost through technology advances.
Addressing the content of the postscript, the Board writes:
In the postscript to P-1572, Assistant Commissioner Mitchinson speculated, with respect to the ONBIS database, that the cost to have incorporated the design features into the original ONBIS system that would have allowed the data at issue in that case to be extracted would have been considerably lower than the costs to modify the system later to add this functionality. Assistant Commissioner Mitchinson suggests that not doing so indicated a lack of foresight on the part of the Government. … However, as I indicated above, it is often not possible, even with foresight, to obtain an ideal technological system with limitless functionality. Choices and compromises must be made, as was the case when the [Ontario Rental Housing Tribunal] acquired the Caseload software.
The appellant’s representations
The appellant provided extensive technical submissions challenging the assertions of the Board. In his initial submissions he wrote:
… The Board provided no evidence that an independent third party examined the Board computer and concluded that it was not feasible to modify the program to allow searches by key worlds (sic) a fundamental and basis (sic) programming technique that has been around for more than forty years! [Emphasis in original]
Based on his thirty years of work with computer programs, in his opinion resolving what he views as the Board’s search limitations is a very simple matter.
The appellant submits that Order P-1572 does not assist the Board, and submits that the postscript to that order clearly supports his request. He also relies on the postscript to Order PO-2511, submitting that the technical process used by some administrative tribunals to draft their orders has a major impact on the public’s right to access their “body of case law”.
The appellant also relies on former Assistant Commissioner Mitchinson’s discussion in Order PO-2265 regarding the importance of having tribunals operate in an open and transparent manner, in support of the following proposition:
It follows that one could state with absolute certainty that the Parliament intended for the public to have complete and unrestricted access to the administrative Board/Tribunal “body of case law” that by its generic nature should apply the law that was within the public domain.
No Board/Tribunal could “operate in a transparent fashion” if its “body of case law” is hidden from the public by a Government agency that locked the decision in a “vault” and secured it with a password. In my specific circumstances, the Board’s password is the order’s number. …
I believe that [the Board’s Freedom of Information Co-ordinator] misunderstood Mr. Mitchinson in P-1572. His postscript comments revealed that he was uncomfortable with his “technical findings” and gave the Minister a silent but implied warning that, in the future, the “lack of foresight” defense would not be allowed in similar circumstances. …
In sur-reply, the appellant writes:
I do not question the Board’s right to decide what computer software to use for its operations and that was not the purpose of my request to access its orders. My interest lies exclusively with the body of case law within the Board’s decisions, which content could only be accessed if these documents were searchable by keywords.
The appellant then sets out that he carefully reviewed the content of the memorandum provided by the Board’s Applications Services Unit Manager. It appeared to the appellant that the Board’s Applications Services Unit Manager was not aware that the Ministry of Government Services (MGS) has all the technological expertise and tools required to perform keywords searches of orders. His review of the memorandum led him to infer the following:
The Board’s software stores all its data on MGS file servers in three functional directories: (a) Case Management database; (b) Order Documents; (c) Hearing recording documents (audio files).
Caseload database stores at least two pointers (links) containing the file name of the Order issued and the file name of the hearing’s recording in each specific case adjudicated by the Board (this is the standard implementation of most case management applications, as documents are not stored within the database itself but outside its repository).
The Board orders are most likely grouped in one directory that contains a various number of subdirectories that are probably listed by the year of the order (these orders are created by the Board’s Members in one of the standard word processing program, namely Microsoft Word, Word Perfect or Adobe Acrobat).
…
[The Board’s Applications Services Unit Manager] indicated in his memo that the Board’s orders are “documents stored on a file server”. As such, a most basic search engine should be able to “string search” any order by keywords. MGS has and employs such a search engine in their main web page. [webpage reference omitted].
The same engine provides an even more comprehensive “Advance Search” allowing the user to filter the search results to limit the number of responding hits. [webpage reference omitted].
The appellant then refers to other MGS groups using this search engine or its derivative to search documents by keywords and he appellant then offers another solution to searching the orders by keyword, pointing to publicly available desktop search engines. He writes:
Three companies provide free Desktop Search Engines that could search the contents of documents created in about 200 different formats. [The appellant then lists the companies, which include Microsoft] … Several other companies provide more advanced search engines at higher costs.
In her response of September 19, 2008, [the Board’s Coordinator] calculated that the Board issued 172,259 orders since 1998. Such precise number implies that the Board could account for this large number of orders stored on the MGS file servers.
The appellant also outlines particular courses of action that, he submits, would satisfy his request. He explains that, based on the figures provided by the Board’s Coordinator:
… a very simple calculation would show that every Board member and staff could have had access at his/her desktop computer to all the orders issued in the last ten years.
In order to estimate the size of the server directory required to save all the Board’s orders since 1998, I would assume that:
• Number of orders = 200,000
• The average order takes two pages in a word processor
• The order’s file size is 40KB (that is the size generated by Microsoft Word)
Consequently, the total disk space required is about 8GB, which could easily fit in two rewritable DVDs or one super DVD. Any Board member/staff could then install on their computers one of the free Desktop Search Engines, let it index all the related documents and then access the order by keywords at will.
The appellant also suggests that if the Board is not prepared to do this, it could supply him the data on a DVD and he would do the work using his own computer. The appellant states that would be prepared to sign any required confidentiality agreement.
The Board’s response to the appellant’s solutions
The Board submits in response that the solutions suggested by the appellant are fully addressed in the memorandum from its Applications Services Unit Manager. In particular, it submits that:
To reiterate, Board orders are not stored and indexed in a way that allows for a key word search. Board orders are written in Microsoft Word and are stored in a very basic manner on a very large file server. They are not stored in a directory with various indexed subdirectories, as the appellant has wrongly assumed … .
The only way to retrieve specific orders from the file server on which they are stored is through a Caseload search. Caseload can only search for specific orders based on the case management information that it stores. Since caseload does not store information about the contents of an order, Caseload does not “know” what words a Board member used when they wrote their order, and it cannot search for an order based on those words.
…
The suggestions for alternate solutions provided in the appellant’s representations are also dealt with in the [memorandum (incorrectly referred to as an affidavit in the Board’s submissions)] - the submission describes metadata document imaging solutions, which, as set out in the [memorandum] requires technical expertise and resources the Board does not have. To be able to apply the so-called “free” desktop search engines described in the appellant’s submission, the Board’s orders would have to already be stored in structured, indexed directories, and doing this is in itself a major IT project.
The appellant’s reply to the Board’s response to his solutions
The appellant submits that the Board’s technical submissions indicate a “total lack of basic understanding on how computer files are stored and retrieved.” The appellant provides the following examples:
At page 1, paragraph 2, the Board stated: “Board orders are written in Microsoft Word and are stored in a very basic manner on a very large file server.”
If the Board orders are written in Microsoft Word, as I suspected, and are stored in a very “basic manner” then a very basic and free search engine could “crawl” through the entire content of these files and detect any single keyword and/or sequence of keywords. This kind of computer knowledge is elementary, and, assuming that MGS’s servers did not possess the magical properties of a “black hole”, it was difficult to believe that a competent MGS engineer would have denied these fundamental facts. [Emphasis in original]
At page 2, paragraph 3, the Board stated:
“The suggestions for alternate solutions provided in the appellant’s representations are also dealt with in the affidavit - the submission describes metadata document imaging solutions, which, as set out in the affidavit requires technical expertise and resources the Board does not have. To be able to apply the so-called “free” desktop search engines described in the appellant’s submission, the Board’s orders would have to already be stored in structured, indexed directories, and doing this is in itself a major IT task.”
The two sentences in this paragraph referred to two unrelated matters:
(i) The “affidavit” speculated briefly (one sentence only) and casually of a possible alternative implementation of a “metadata document imaging solution”. And this was nothing more than [the Board’s Applications Services Unit Manager] dreaming of far-off and impractical solutions that had nothing to do with the matter under appeal. While MGS might want to build a space shuttle to cross a busy street, a reasonable person would achieve the same result by walking when the stop-light turns green, at no additional cost to the tax payer.
(ii) The reference to the “free desktop search engine” was perhaps the best indication that the Board, or whoever wrote the submission, had zero understanding of how computers store and retrieve files.
I indicated in my submission that the “most likely” way to save files was to create a directory tree with a limited number of files in its subdirectories. Most computer engineers and users know this basic implementation of data storage. lt was hard to believe that MGS chose to cram 200,000 files in one directory but, even if they did so, it had no functional relevance on a search engine: The engine’s indexing and retrieval of content by keywords is exactly the same for one directory or for an extensive subdirectory tree. While the search response time might vary, the
results would be identical! [Emphasis in original]
The Board’s Current Search Engine
As set out above, the ability of the Board to accommodate requests for searches of its decisions changed over the course of the appeal. The Board explained in its final submission that it is now redacting personal information from contested orders as they are issued and posting them in the Redacted Orders section of the website. The Board advised that it will take some time before all the full range of contested orders issued under the RTA are available. The Board advised that it will not be redacting orders issued by the ORHT under the Tenant Protection Act.
The Board wrote:
In the Board’s representations for this appeal, we had set out that the Board orders are not stored and indexed in a way that allows for a keyword search. However, it is now possible to search the redacted orders on the website using key words and the website’s search feature. There are some limitations, however. For example, the search feature will also find documents from other websites that contain the terms used in the search. Searching the redacted orders based on the term “mortgagee” yielded 766 “hits”, of which a dozen or so were Board orders. Also, the Board’s orders are not always found on the first page of the search results. For example, when the term “Mortgages Act” was used, the first Board order included in the search results was found on the second page of the results. However, Board orders are identified in the search results based on the file number (for example, “TEL-12345”), and it is not difficult to find them.
While it is still not possible to provide the appellant with every order ever issued by the Board related to the key words found in his original request (and, as indicated above, the Board will not be redacting orders issued under the TPA), a number of orders related to the key words identified by the appellant are available, and additional orders will be available in the future as the Board continues to redact orders previously issued under the RTA.
I regret that I was unable to provide information about this feature at the time I wrote earlier representations for this appeal. However, at that time, the Redacted Orders were not yet available on the Board’s website. As well, the Board did not yet know whether the search feature available on the Board’s website would be effective for searching for redacted orders by key words. As a result, I was not in a position to refer to the Redacted Orders feature of the website in my representations.
The appellant welcomed the initiative, but was not satisfied with it. He also took the position that in the circumstances, the Board’s previous submissions should be disregarded and his be adopted in their entirety. He wrote:
The good part of the Board’s decision to disclose its orders to the public was the fact that the orders are now redacted one by one and made available online. And this is the bulk of the task ahead. However, without a proper and efficient display of the search results, the overall effect would be less than satisfactory. All it takes now is for MGS to assign the task of designing the search engine to an experienced engineer that understands the technology, which is within the main stream of computer knowledge.
The appellant further submits that he conducted a search in the manner suggested by the Board’s Coordinator, but the results were unsatisfactory. He noted:
When I entered the term “mortgagee in possession” (with quotation marks) the search engine returned 63 hits listed on 7 pages (10 hits per page), each result containing this sentence correctly. Unfortunately, out of 63 hits, only about 10 were from the Board’s decisions. The rest of the search results were from other government and legal documents created elsewhere.
My conclusion of the new search options was that the Board decided to finally open the faucet allowing its orders to flow out freely but it provided the user with a spaghetti strainer to drink the water (decisions).
He then provided a number of “important factors” that the IPC should consider in any order it may make in the appeal, which should cover all Board decisions from 1998 and also include continued monitoring of the Board until it completes the “redaction” of all of its previous orders.
The appellant’s final search
The last submission that the appellant provided in this appeal related to his subsequent review of the Board’s website. He wrote:
The most important section of law covers the Board’s orders to evict tenants following a landlord application using Forms L1 - L4. Under these four order types, [the Board] listed some of the orders released during five months only, Jan - Mar 2009 and Feb - Mar 2007. The Board provided no advice when it would complete listing all its previous orders for the past ten years and especially the most recent decisions issued during the last 2 years, 2008/9. The Board’s listing of other types of orders was even less.
…
At this “in progress” pace (the slogan used to explain unlisted orders), the Board would never catch up with the old orders as new ones are constantly released at higher than ever pace, due to the severe economic conditions of our present times.
Analysis and Finding
As noted earlier, the issue before me in this appeal is whether the Board’s fee estimate should be upheld. In this case, however, the issue also requires consideration of the Board’s technological approach to the request, which is the basis for the fee, and whether that proposed approach is reasonable. Thus the context of the request, the nature of the records and the manner in which the requested information could be located using technological solutions must be taken into account.
In doing so, a number of sections of the Act and Regulation 460 have potential relevance. I will begin by setting out these provisions.
Section 1 of the Act sets out that the purposes of the Act are,
(a) to provide a right of access to information under the control of institutions in accordance with the principles that,
(i) information should be available to the public,
(ii) necessary exemptions from the right of access should be limited and specific, and
(iii) decisions on the disclosure of government information should be reviewed independently of government; and
(b) to protect the privacy of individuals with respect to personal information about themselves held by institutions and to provide individuals with a right of access to that information.
Subsection 10(1) of the Act sets out a person’s general right of access to records:
Every person has a right of access to a record or a part of a record in the custody or under the control of an institution unless,
(a) the record or the part of the record falls within one of the exemptions under sections 12 to 22; or
(b) the head is of the opinion on reasonable grounds that the request to access is frivolous or vexatious
Section 24(1) of the Act imposes certain obligations on requesters and institutions when submitting and responding to requests for access to records. This section states, in part:
(1) A person seeking access to a record shall,
(a) make a request in writing to the institution that the person believes has custody or control of the record;
(b) provide sufficient detail to enable an experienced employee of the institution, upon a reasonable effort, to identify the record;
. . .
(2) If the request does not sufficiently describe the record sought, the institution shall inform the applicant of the defect and shall offer assistance in reformulating the request so as to comply with subsection (1).
Section 57(1) requires an institution to charge fees for requests under the Act. That section reads:
A head shall require the person who makes a request for access to a record to pay fees in the amounts prescribed by the regulations for,
(a) the costs of every hour of manual search required to locate a record;
(b) the costs of preparing the record for disclosure;
(c) computer and other costs incurred in locating, retrieving, processing and copying a record;
(d) shipping costs; and
(e) any other costs incurred in responding to a request for access to a record.
More specific provisions regarding fees are found in sections 6, 6.1, 7 and 9 of Regulation 460. Section 6(5) of Regulation 460 reads:
The following are the fees that shall be charged for the purposes of subsection 57(1) of the Act for access to a record:
For developing a computer program or other method of producing a record from machine readable record, $15 for each 15 minutes spent by any person.
Section 2(1) of the Act specifically defines a “record” as follows:
“record” means any record of information however recorded, whether in printed form, on film, by electronic means or otherwise, and includes,
(a) correspondence, a memorandum, a book, a plan, a map, a drawing, a diagram, a pictorial or graphic work, a photograph, a film, a microfilm, a sound recording, a videotape, a machine readable record, any other documentary material, regardless of physical form or characteristics, and any copy thereof, and
(b) subject to the regulations, any record that is capable of being produced from a machine readable record under the control of an institution by means of computer hardware and software or any other information storage equipment and technical expertise normally used by the institution;
Section 2 of Regulation 460 under the Act states:
A record capable of being produced from machine readable records is not included in the definition of “record” for the purposes of the Act if the process of producing it would unreasonably interfere with the operations of an institution.
This office has previously stated that government organizations are not obliged to maintain records in such a manner as to accommodate the various ways in which a request for information might be framed [See the postscript to Order M-583]. However, this office has also stated that institutions have an obligation to maintain their electronic records in formats that ensure expeditious access and disclosure in a manner or form that is accessible by all members of the public. In the electronic age, this is essential for an open and transparent government institution. [See Order MO-2199]. Furthermore, in the postscript to Order P-1572, former Assistant Commissioner Mitchinson emphasized that as parts of government become increasingly reliant on electronic databases to deliver their programs, it is critically important that public accessibility considerations be part of the decision-making process on any new systems design.
As set out above, the term “record” is defined in section 2(1) of the Act. Paragraph (b) of the definition provides that subject to the regulations, the term “record” includes any record that is capable of being produced from a machine readable record under the control of an institution by means of computer hardware and software or any other information storage equipment and technical expertise normally used by the institution.
As explained by Adjudicator Colin Bhattachargee in Order MO-2129, the term, “machine readable record,” is not defined in the Act. However, a machine readable record can be defined as a record that is capable of being rendered intelligible by a machine. For example, a machine readable record would include a database that is capable of being rendered intelligible by a computer. Other examples of machine readable records would include a DVD which can be played or rendered intelligible by a DVD player or an audiotape which can be listened to or rendered intelligible by a tape recorder.
Section 2 of Regulation 460 under the Act (quoted above) places further limits on the definition of “machine readable record” in section 2 of the Act by excluding a record from that definition “if the process of producing it would unreasonably interfere with the operations of an institution.”
In Order P-50, Former Commissioner Linden stated that what constitutes “unreasonable interference” is a matter which must be considered on a case-by-case basis, but it is clear that the regulation is intended to impose limits on the institution’s responsibility to create a new record.
Moreover, section 6(5) of the same regulation provides for a fee to be charged by an institution “for developing a computer program or other method of producing a record from a machine readable record.”
In Order MO-2129 Adjudicator Bhattacharjee addressed a request for information that appeared to exist within the record holdings of an institution, but not in the format asked for by the appellant in that appeal. In particular, the Toronto Police Services Board (the Police) submitted that the requested information did not exist in list format and that their database does not contain a report function capable of extracting a list of agencies. However, they acknowledged that the information that could be used to compile a list exists in both paper and electronic records.
In finding that the Police failed to adequately respond to the request in that appeal, Adjudicator Bhattacharjee wrote:
I agree with former Commissioner Linden’s reasoning in Order P-50. Clearly, it would be unreasonable to expect an institution to create a record in the format sought by the requester if the records are the type contemplated by paragraph (a) of the definition of a record, such as paper records. Unless the records are few in number, it would require significant resources and staff time for an institution to manually organize the data from such records into the format sought by the requester, such as a list.
However, there is a clear policy rationale underlying the special rules governing computerized or electronic records inherent in paragraph (b) of the definition of a record. The data in a machine readable record, such as a database, can be retrieved, manipulated and reorganized with ease through the use of information technology tools, such as computer software. Consequently, in comparison to paper records, it is significantly easier and less labour intensive for institutions to organize electronic data into the format sought by the requester. This is why section 2(1) of the Act defines a record as including any record that is capable of being produced from a machine readable record in the circumstances set out in paragraph (b).
In 1997, this office published a paper, Electronic Records: Maximizing Best Practices, that provided further commentary on this requirement:
The Acts and regulations recognize the obligation of government organizations to create electronic records when requested, except where to do so would unreasonably interfere with the operations of the government organization. That obligation would be satisfied through the use of the appropriate hardware and software to create the document …
Adjudicator Bhattacharjee went on to address the obligations of the Police in the circumstances of that appeal, determining that:
… If the request is for information that currently exists in a recorded format different from the format asked for by the requester, as is the case in this appeal, the Police have dual obligations.
First, if the requested information falls within paragraph (a) of the definition of a record (e.g., paper records), the Police have a duty to identify and advise the requester of the existence of these related records (i.e., the raw material). However, the Police are not required to create a record from these records that is in the format asked for by the requester (e.g., a list).
Second, if the requested information falls within paragraph (b) of the definition of a record, the Police have a duty to provide it in the requested format (e.g., a list) if it can be produced from an existing machine readable record (e.g., a database) by means of computer hardware and software or any other information storage equipment and technical expertise normally used by the institution, and doing so will not unreasonably interfere with the operations of the Police. In such circumstances, the Police have a duty to create a record in the format asked for by the requester.
In my view, a reasonable search for records responsive to an access request would include taking steps to comply with these two obligations. …
In Toronto Police Services Board v. (Ontario) Information and Privacy Commissioner, 2009 ONCA 20 (CanLII), the Ontario Court of Appeal discussed the application of a contextual and purposive analysis of the identically worded section 2(1)(b) of the Municipal Freedom of Information and Protection of Privacy Act:
A contextual and purposive analysis of s. 2(1)(b) must also take into account the prevalence of computers in our society and their use by government institutions as the primary means by which records are kept and information is stored. This technological reality tells against an interpretation of s. 2(1)(b) that would minimize rather than maximize the public’s right of access to electronically recorded information.
The Divisional Court made no mention of these principles of interpretation in constructing s. 2(1)(b) of the Act and in concluding that the Adjudicator’s interpretation was unreasonable. This omission led the court to give s. 2(1)(b) a narrow construction – one which, in my respectful view, fails to reflect the purpose and spirit of the Act and the generous approach to access contemplated by it.
The Divisional Court’s interpretation of s. 2(1)(b) would eliminate all access to electronically recorded information stored in an institution’s existing computer software where its production would require the development of an algorithm or software within its available technical expertise to create and using software it currently has. In my view, other provisions in the Act and the regulations tell against this interpretation.
Sections 45(1)(b) and (c) of the Act require the requester to bear the “costs of preparing the record for disclosure” and “computer and other costs incurred in locating, retrieving, processing and copying a record,” in accordance with the fees prescribed by the regulations. Subsections 6(5) and (6) of Reg. 823 were enacted pursuant to s. 45(1) of the Act. These provisions state:
6. The following are the fees that shall be charged for the purposes of subsection 45(1) of the Act for access to a record:
…
5. For developing a computer program or other method of producing a record from machine readable record, $15 for each 15 minutes spent by any person.
6. The costs, including computer costs, that the institution incurs in locating, retrieving, processing and copying the record if those costs are specified in an invoice that the institution has received.
In my view, a liberal and purposive interpretation of those regulations when read in conjunction with s. 2(1)(b), which opens with the phrase “subject to the regulations,” and in conjunction with s. 45(1), strongly supports the contention that the legislature contemplated precisely the situation that has arisen in this case. In some circumstances, new computer programs will have to be developed, using the institution’s available technical expertise and existing software, to produce a record from a machine readable record, with the requester being held accountable for the costs incurred in developing it. [reference omitted]
While I can appreciate that technology evolves over time, in its representations the Board acknowledges that its “orders are written in Microsoft Word and are stored in a very basic manner on a very large file server”. To refute the appellant’s assertion that publicly available search engines could search this file server, the Board states that in order to apply the desktop search engines, the orders would have to be stored in structured indexed directories “which is, in itself, a major IT project.” The appellant challenges this assertion submitting that “a very basic and free search engine [which as I noted above, included a Microsoft product] could “crawl” through the entire content of these files and detect any single keyword and/or sequence of keywords.” As this type of search is based on the keywords sought by the appellant, it would also negate the need for the two-step process suggested by the Board: identifying the type of application that would likely contain a keyword and then reviewing the associated decision to see if a keyword was contained in it. As set out in the Board’s fee estimate and its representations, this two step process resulted in the search fee of $86,130.00. The estimated fee for severing responsive records was only $344.00.
The memorandum of the Board’s Application Services Manager suggests a cost of $75,000.00 to $100,000 and an elapsed time of 6-9 months to formulate a complex metadata solution as an alternative to the two step manual search set out by the Board in its fee estimate and its representations. At the same time it does not provide a fee estimate for a much simpler database search of the very large file server along the lines first suggested by the appellant in his sur-reply representations that were shared with the Board.
The Board’s initial position was that conducting a search for records in its custody and control would be costly and time consuming, and that a search engine by keyword or term was beyond its capacity. Over time it revealed how the decisions were actually stored. Its representations culminated in revealing the ongoing development of a public, web based search engine to which its orders are gradually being added over time. The ORHT’s orders, issued under the former Tenant Protection Act, will not be included. This most recent solution does not address the appellant’s request because the database is incomplete and does not include ORHT decisions.
I have carefully considered the evidence and argument provided to me in this case. In my view, the Board’s fee for searching would not be reasonable and should not be upheld if it is based on a costly technological solution where inexpensive alternative solutions are readily available. While I am not in a position to order the Board to use a particular type of technology that it has not previously used, I am also not able to uphold the unreasonable use of an expensive technological solution when the evidence supports the ready availability of an inexpensive or free solution to the same problem.
On the evidence presented, I conclude that the approach taken by the Board, in light of publicly available search technology and the fact that the decisions to be searched are “are written in Microsoft Word and are stored in a very basic manner on a very large file server”, is unreasonable in the circumstances. I prefer the evidence of the appellant, who has, in my view, demonstrated the existence of readily available, inexpensive solutions. The appellant should not be penalized in fees for the Board’s failure to use these solutions. Accordingly, I will disallow the search component of the Board’s fee estimate in the sum of $86,130.00.
The Board intends to sever the records, and I find its fee estimate in that regard, and its proposed photocopy charges, to be reasonable. I uphold the Board’s estimated fee of $344.00 in record preparation time as well as photocopy charges of $137.60. The photocopy charges and the Board’s estimated fee for record preparation time can be adjusted if the actual number of copies or the fee for record preparation time differs from the Board’s projection.
ORDER:
1. I uphold the Board’s fee estimate only to the extent of $344.00 in record preparation time and photocopy charges of $137.60.
2. I order the Board to issue a final access decision to the appellant, with a copy to me, no later than September 14, 2010.
_____________________________________ August 13, 2010