Is Metadata a Public Record? Part 2
Published: 03/18/10
Author Name: Kara Millonzi
In my last post I defined metadata and summarized the three cases (in other states) that have addressed whether, and to what extent, metadata is a public record. I concluded that post by asking what we might glean from these cases to determine how North Carolina’s public records laws apply to metadata. The simple answer to that question is “not much.” Those cases involved interpretations of different public records laws and each addressed different types of metadata. And, not surprisingly, the three courts arrived at somewhat different conclusions with respect to the application of their states’ public record laws to the metadata at issue. Perhaps the only useful thing to note is that all three courts determined that at least some metadata constituted a public record and, as such, was subject to public disclosure, at least when the metadata was specifically requested.
Having said that, I would like to reframe the inquiry for purposes of this post to address the following two questions: (1) Under North Carolina law, what, if any, metadata is part of a public record that exists in an electronic form? and (2) Under North Carolina law, what does it mean for metadata to be part of a public record for purposes of public access to that metadata?
What, if any, metadata is part of an electronic public record?
It is important to emphasize at the outset that there is no North Carolina case law addressing this issue (yet). However, given the state’s broad definition of public records it is likely that at least some metadata constitutes a public record in North Carolina. G.S. 132-1 defines public records to include “all . . . electronic data-processing records . . . regardless of physical form or characteristics . . . made or received pursuant to law or ordinance in connection with the transaction of public business.”
The following are three possible approaches that North Carolina courts might take in answering the question of what metadata is part of an electronic public record.
1. All metadata associated with an electronic public record is part of that public record.
This approach has the advantage of being straight-forward, at least from a legal perspective. And, it arguably is consistent both with the plain language of North Carolina’s public records statutes and with the broad interpretation afforded that statutory language by North Carolina courts.
As a practical matter, however, it may be very difficult to provide the public with access to all of the metadata associated with a public record. Recall that the “definition” of metadata also is very broad. It is not limited to information embedded within a document or file. It also encompasses system profile information about that document or file; information that a computer system uses to track and retrieve the document or file; and information about the relationship of a particular document or file to other documents or files within the computer network (system metadata)—all of which are not embedded within a particular record. A person likely does not have a right to access a public agency’s active computer system in order to view all of a record’s system metadata. See G.S. 132-6.1(c). However, some metadata may not be viewable without this access. Furthermore, some metadata is highly volatile and may be altered even by saving an electronic record to a disk or forwarding an e-mail. Thus, it may be difficult, if not impossible, for a public agency to provide access to, or a copy of, all of the metadata associated with a particular public record
2. Only the metadata associated with an electronic public record that falls into one or more of three “metadata categories” is part of that public record.
For legal or practical reasons (or both) North Carolina courts may determine that not all metadata constitutes a public record. And the courts may borrow from the emerging case law in the civil discovery context to inform what metadata associated with a public record is part of that public record. As summarized in my previous blog post, several courts have adopted a categorical approach to aid in determining when, and to what extent, metadata is discoverable in civil litigation. Under this case law, metadata generally is grouped into one of three categories—substantive/application metadata (e.g., fonts, spacing, edit history), system metadata (e.g., date created, date last accessed, date last modified, author, searchability and retrieval functions), and embedded metadata (e.g., spreadsheet cell formulae, hyperlinks, externally linked files).
This approach explicitly acknowledges that not all metadata is created equally. Not surprisingly, two of the courts that have addressed the application of their state’s public records laws to metadata have referenced these categorical distinctions and at least suggested that only metadata in a certain category constitutes a public record subject to public disclosure. The courts in the two cases, however, came to different conclusions about which category of metadata was subject to public access under their respective state’s public records laws.
This illustrates the potential difficulty in applying the categorical distinctions in the public records context. These categories may be useful in determining whether metadata is discoverable in litigation—an inquiry that is often done on a case-by-case basis. One or more categories of metadata may be discoverable in one case, whereas a different category of metadata may be discoverable in a different case. The categorical distinctions may not work as well when applied to public records, where access is not dependent on a person’s purpose in requesting a record. Furthermore, it is important to provide public agencies with clear rules that apply to all public records.
3. Only the metadata associated with a public record that is “essential” or “integral” to the record is part of that public record.
In some respects, this approach is a hybrid of the first two, and it may make the most sense for a practical application of North Carolina’s public records laws. It does not define a public record to include all of its associated metadata, but it also does not create categorical distinctions based only on whether or not the metadata is embedded within the electronic record. Instead, it looks at the value of the metadata to the record with which it is associated to frame the analysis. Some metadata is required to comprehend the underlying record. Other metadata is important to understanding the context of the underlying record. But some metadata simply provides no useful information about the underlying record as a public record.
The difficulty with this approach, of course, is determining exactly what metadata is “essential” or “integral” to the underlying record. There are a few easy distinctions, though. Some system metadata, for example, such as e-mail header information or the date a record was last modified and by whom, may be vital to fully understanding the context of a record or demonstrating the record’s integrity. It also may be easily analogized to information that may exist on a paper document. Other system metadata, such as that which aids a computer in retrieving a file or document, may have little to no bearing on the record as a public record. Along that line, much of the metadata categorized as embedded metadata likely is essential or integral to a public record. For example, it often is difficult, if not impossible, to understand the data that appears on the face of a spreadsheet without access to the hidden formulae. On the other hand, some of the metadata categorized as substantive/application metadata may not be essential or integral to the record, even though it also technically is embedded within the document or file.
It bears repeating that these are only suggested potential approaches to the application of North Carolina public records laws to metadata. And, none of the proposed approaches address all of the legal and practical challenges related to determining whether, and to what extent, metadata constitutes a public record. In the absence of legislative or judicial guidance, a public agency’s employees should consult with the agency’s legal counsel to determine how metadata should be treated in the public records context.
Note also that none of the proposed approaches address the issue of the retention of metadata associated with a public record. The North Carolina Department of Cultural Resources (NCDCR) prescribes retention and disposition requirements for public records. To date, the state’s retention schedules do not specifically address the retention and disposition of metadata associated with a public record. Thus, it is unclear whether, and to what extent, a public record’s metadata must be retained according to the retention schedule that applies to that record. I understand that guidance from the NCDCR specifically related to the retention of metadata may be forthcoming. Meanwhile, it may be up to the courts to resolve the question by determining whether there is some metadata that is inherently part of a record, such that it must be preserved in accordance with existing retention rules.
What does it mean for metadata to be part of a public record for purposes of public access to that metadata?
The answer to this question might seem obvious. If metadata is part of record that is subject to public access then that metadata also is subject to public access. And, in fact, that may very well be what North Carolina law requires. G.S. 132-6(a) specifically states that “every custodian of public records shall permit any record in the custodian’s custody to be inspected . . . and shall, as promptly as possible, furnish copies thereof . . . .” And, G.S. 132-6.2(a) states that persons requesting copies of public records “may elect to obtain [the records] in any and all media in which the public agency is capable of providing them.” Presumably then, if a public agency receives a request for a printed copy of a public record that exists in electronic form, it also must provide a printed copy of all the metadata that is part of that public record.
To date, however, no court has held that the default production of a public record must include its associated metadata. In fact, in O’Neill v. City of Shoreline, 187 P.3d 822 (2008), the Washington Court of Appeals indicated that the fact that the plaintiff had specifically requested a copy of the public record (e-mail) in electronic form, and with its associated metadata, was significant to the court’s determination that the metadata at issue (e-mail header information) was subject to public disclosure. The court at least suggested that the defendant-city might not have had to produce the metadata if the request has been for a hard-copy of the e-mail.
Perhaps significantly, at least under North Carolina law, in order to provide a copy of the metadata that is part of a public record in hard-copy form, a local government arguably would have to compile or create a separate record. This is particularly true if a public agency must capture any of a public record’s system metadata. G.S. 132-6.2(e) provides, however, that a public agency is not required to “respond to a request for a copy of a public record by creating or compiling a record that does not exist.” Thus, a public agency in North Carolina might be justified in providing a copy of the metadata that is part of a public record only upon a request for an electronic copy of the public record. (Note that a public agency still would have to permit inspection of the metadata that is part of a public record.) It is unclear how North Carolina courts might reconcile these statutory provisions in this context.
1
Coates’ Canons NC Local Government Law
Is Metadata a Public Record? Part 2
Published: 03/18/10
Author Name: Kara Millonzi
In my last post I defined metadata and summarized the three cases (in other states) that have addressed whether, and to what extent, metadata is a public record. I concluded that post by asking what we might glean from these cases to determine how North Carolina’s public records laws apply to metadata. The simple answer to that question is “not much.” Those cases involved interpretations of different public records laws and each addressed different types of metadata. And, not surprisingly, the three courts arrived at somewhat different conclusions with respect to the application of their states’ public record laws to the metadata at issue. Perhaps the only useful thing to note is that all three courts determined that at least some metadata constituted a public record and, as such, was subject to public disclosure, at least when the metadata was specifically requested.
Having said that, I would like to reframe the inquiry for purposes of this post to address the following two questions: (1) Under North Carolina law, what, if any, metadata is part of a public record that exists in an electronic form? and (2) Under North Carolina law, what does it mean for metadata to be part of a public record for purposes of public access to that metadata?
What, if any, metadata is part of an electronic public record?
It is important to emphasize at the outset that there is no North Carolina case law addressing this issue (yet). However, given the state’s broad definition of public records it is likely that at least some metadata constitutes a public record in North Carolina. G.S. 132-1 defines public records to include “all . . . electronic data-processing records . . . regardless of physical form or characteristics . . . made or received pursuant to law or ordinance in connection with the transaction of public business.”
The following are three possible approaches that North Carolina courts might take in answering the question of what metadata is part of an electronic public record.
1. All metadata associated with an electronic public record is part of that public record.
This approach has the advantage of being straight-forward, at least from a legal perspective. And, it arguably is consistent both with the plain language of North Carolina’s public records statutes and with the broad interpretation afforded that statutory language by North Carolina courts.
As a practical matter, however, it may be very difficult to provide the public with access to all of the metadata associated with a public record. Recall that the “definition” of metadata also is very broad. It is not limited to information embedded within a document or file. It also encompasses system profile information about that document or file; information that a computer system uses to track and retrieve the document or file; and information about the relationship of a particular document or file to other documents or files within the computer network (system metadata)—all of which are not embedded within a particular record. A person likely does not have a right to access a public agency’s active computer system in order to view all of a record’s system metadata. See G.S. 132-6.1(c). However, some metadata may not be viewable without this access. Furthermore, some metadata is highly volatile and may be altered even by saving an electronic record to a disk or forwarding an e-mail. Thus, it may be difficult, if not impossible, for a public agency to provide access to, or a copy of, all of the metadata associated with a particular public record
2. Only the metadata associated with an electronic public record that falls into one or more of three “metadata categories” is part of that public record.
For legal or practical reasons (or both) North Carolina courts may determine that not all metadata constitutes a public record. And the courts may borrow from the emerging case law in the civil discovery context to inform what metadata associated with a public record is part of that public record. As summarized in my previous blog post, several courts have adopted a categorical approach to aid in determining when, and to what extent, metadata is discoverable in civil litigation. Under this case law, metadata generally is grouped into one of three categories—substantive/application metadata (e.g., fonts, spacing, edit history), system metadata (e.g., date created, date last accessed, date last modified, author, searchability and retrieval functions), and embedded metadata (e.g., spreadsheet cell formulae, hyperlinks, externally linked files).
This approach explicitly acknowledges that not all metadata is created equally. Not surprisingly, two of the courts that have addressed the application of their state’s public records laws to metadata have referenced these categorical distinctions and at least suggested that only metadata in a certain category constitutes a public record subject to public disclosure. The courts in the two cases, however, came to different conclusions about which category of metadata was subject to public access under their respective state’s public records laws.
This illustrates the potential difficulty in applying the categorical distinctions in the public records context. These categories may be useful in determining whether metadata is discoverable in litigation—an inquiry that is often done on a case-by-case basis. One or more categories of metadata may be discoverable in one case, whereas a different category of metadata may be discoverable in a different case. The categorical distinctions may not work as well when applied to public records, where access is not dependent on a person’s purpose in requesting a record. Furthermore, it is important to provide public agencies with clear rules that apply to all public records.
3. Only the metadata associated with a public record that is “essential” or “integral” to the record is part of that public record.
In some respects, this approach is a hybrid of the first two, and it may make the most sense for a practical application of North Carolina’s public records laws. It does not define a public record to include all of its associated metadata, but it also does not create categorical distinctions based only on whether or not the metadata is embedded within the electronic record. Instead, it looks at the value of the metadata to the record with which it is associated to frame the analysis. Some metadata is required to comprehend the underlying record. Other metadata is important to understanding the context of the underlying record. But some metadata simply provides no useful information about the underlying record as a public record.
The difficulty with this approach, of course, is determining exactly what metadata is “essential” or “integral” to the underlying record. There are a few easy distinctions, though. Some system metadata, for example, such as e-mail header information or the date a record was last modified and by whom, may be vital to fully understanding the context of a record or demonstrating the record’s integrity. It also may be easily analogized to information that may exist on a paper document. Other system metadata, such as that which aids a computer in retrieving a file or document, may have little to no bearing on the record as a public record. Along that line, much of the metadata categorized as embedded metadata likely is essential or integral to a public record. For example, it often is difficult, if not impossible, to understand the data that appears on the face of a spreadsheet without access to the hidden formulae. On the other hand, some of the metadata categorized as substantive/application metadata may not be essential or integral to the record, even though it also technically is embedded within the document or file.
It bears repeating that these are only suggested potential approaches to the application of North Carolina public records laws to metadata. And, none of the proposed approaches address all of the legal and practical challenges related to determining whether, and to what extent, metadata constitutes a public record. In the absence of legislative or judicial guidance, a public agency’s employees should consult with the agency’s legal counsel to determine how metadata should be treated in the public records context.
Note also that none of the proposed approaches address the issue of the retention of metadata associated with a public record. The North Carolina Department of Cultural Resources (NCDCR) prescribes retention and disposition requirements for public records. To date, the state’s retention schedules do not specifically address the retention and disposition of metadata associated with a public record. Thus, it is unclear whether, and to what extent, a public record’s metadata must be retained according to the retention schedule that applies to that record. I understand that guidance from the NCDCR specifically related to the retention of metadata may be forthcoming. Meanwhile, it may be up to the courts to resolve the question by determining whether there is some metadata that is inherently part of a record, such that it must be preserved in accordance with existing retention rules.
What does it mean for metadata to be part of a public record for purposes of public access to that metadata?
The answer to this question might seem obvious. If metadata is part of record that is subject to public access then that metadata also is subject to public access. And, in fact, that may very well be what North Carolina law requires. G.S. 132-6(a) specifically states that “every custodian of public records shall permit any record in the custodian’s custody to be inspected . . . and shall, as promptly as possible, furnish copies thereof . . . .” And, G.S. 132-6.2(a) states that persons requesting copies of public records “may elect to obtain [the records] in any and all media in which the public agency is capable of providing them.” Presumably then, if a public agency receives a request for a printed copy of a public record that exists in electronic form, it also must provide a printed copy of all the metadata that is part of that public record.
To date, however, no court has held that the default production of a public record must include its associated metadata. In fact, in O’Neill v. City of Shoreline, 187 P.3d 822 (2008), the Washington Court of Appeals indicated that the fact that the plaintiff had specifically requested a copy of the public record (e-mail) in electronic form, and with its associated metadata, was significant to the court’s determination that the metadata at issue (e-mail header information) was subject to public disclosure. The court at least suggested that the defendant-city might not have had to produce the metadata if the request has been for a hard-copy of the e-mail.
Perhaps significantly, at least under North Carolina law, in order to provide a copy of the metadata that is part of a public record in hard-copy form, a local government arguably would have to compile or create a separate record. This is particularly true if a public agency must capture any of a public record’s system metadata. G.S. 132-6.2(e) provides, however, that a public agency is not required to “respond to a request for a copy of a public record by creating or compiling a record that does not exist.” Thus, a public agency in North Carolina might be justified in providing a copy of the metadata that is part of a public record only upon a request for an electronic copy of the public record. (Note that a public agency still would have to permit inspection of the metadata that is part of a public record.) It is unclear how North Carolina courts might reconcile these statutory provisions in this context.
All rights reserved. This blog post is published and posted online by the School of Government to address issues of interest to government officials. This blog post is for educational and informational use and may be used for those purposes without permission by providing acknowledgment of its source. Use of this blog post for commercial purposes is prohibited. To browse a complete catalog of School of Government publications, please visit the School’s website at www.sog.unc.edu or contact the Bookstore, School of Government, CB# 3330 Knapp-Sanders Building, UNC Chapel Hill, Chapel Hill, NC 27599-3330; e-mail sales@sog.unc.edu; telephone 919.966.4119; or fax 919.962.2707.
One Response to “Is Metadata a Public Record? Part 2”
Washington Supreme Court Cites LTA Journal in Landmark Metadata Opinion « Law, Technology & Arts Blog
[…] of whether metadata constitutes a public record. (Kara Millonzi at Coates’ Canons blog has an excellent analysis of how the law might evolve in North Carolina.) Several useful secondary sources include: Access to […]