Sustainability of Digital Formats: Planning for Library of Congress Collections

Introduction | Sustainability Factors | Content Categories | Format Descriptions | Contact
Format Description Categories >> Browse Alphabetical List

ZIP File Format (PKWARE)

>> Back
Table of Contents
Format Description Properties Explanation of format description terms

Identification and description Explanation of format description terms

Full name ZIP File Format (PKWARE)

This format is designed for cross-platform data exchange and efficient data storage for a set of related files. ZIP_PK is a de facto industry standard, developed, maintained, and openly documented by PKWARE. The original version of the format was developed by Phil Katz (hence the "PK" in PKWARE). ZIP_PK combines data compression, file management, and data encryption within a portable archive format. A ZIP file is a package containing one or more files, usually compressed and sometimes encrypted. The basic structure consists of a sequence of chunks comprising a "local file header" followed by the file data (after compression and/or encryption) followed by a chunk known as the "central directory," which lists the files in the package along with key metadata to support their extraction, decryption, etc. Over the years the format has been extended by PKWARE to adapt to new expectations of users and to changing technology, for example, to allow improved methods for compression and encryption, to support larger file sizes, and to support UNICODE UTF-8 for filenames. See Notes below for more detail on chronological versions.

Over the years a few variants of the ZIP format have been introduced by other companies, variants that were not necessarily fully compatible with any particular ZIP_PK version, with the differences often being related to timing of introduction of new features, such as support for long file names or stronger encryption. For clarity, the compilers of this resource have chosen to use the name "ZIP_PK" for the format described in this document. See Notes below for a brief discussion of variants.

ZIP_PK has been adopted as the package or container for several digital formats that represent a single document or other logical unit but comprise multiple files. These include:

  • Office Open XML (OOXML, standardized as ECMA 376 and ISO/IEC 29500)
  • Office Document Format (ODF, from OASIS, later standardized as ISO/IEC 26300)
  • EPUB (for electronic publications, originally developed by the International Digital Publishing Forum (IDPF) and since 2017 under the auspices of the World Wide Web Consortium (W3C). See EPUB_3_0 (ISO/IEC TS 30135:2014) and EPUB_3_2.
  • Java ARchive (JAR, used to distribute software applications or libraries in the Java programming language).
Production phase May be used at any lifecycle phase for bundling/packaging files together for exchange, storage, or distribution. In particular, XML-based office productivity formats use a ZIP package that may be used during the initial creating phase; and XML-based formats such as EPUB, intended for distributing published (final-stage) documents, also use ZIP as a package or container.
Relationship to other formats
    Has subtype ZIP_6_2_0, ZIP file format, version 6.2.0 (PKWARE). Version 6.2.0 of the APPNOTE.TXT from PKWARE, published in 2004, has been incorporated into Office OpenXML (OOXML, ISO 29500-2:2008-2012) and the Open Document Format (ODF: 1.0 and 1.2 as published by OASIS) as the document container format.
    Has subtype ZIP_6_3_3, ZIP File Format, Version 6.3.3 (PKWARE). Chosen as the basis for a proposed ISO specification (to be ISO/IEC 21320, once approved) for a Document Container Format.
    Has subtype ZIP_21320_1, Document Container File: Core (based on ZIP 6.3.3). ISO/IEC 21320-1 Information technology -- Document Container File -- Part 1: Core
    Has subtype The ZIP_PK format has other chronological versions, more of which may be separately described in the future, for example, if specific versions are incorporated into other specifications.
    Has modified version Other entities have introduced variant format versions using the .zip extension, not strictly compatible with any particular chronological version of ZIP_PK, but using its extension capabilities. See Notes below for a brief discussion of variants and compatibility.
    Has subtype SLPK, SLPK Scene Layer Package (for I3S layers)
    Has subtype KMZ, KML (formerly Keyhole Markup Language) Zipped
    Has subtype USDZ, Universal Scene Description Zipped (USDZ)
    Has subtype WACZ, Web Archive Collection Zipped
    Used by 3MF, 3D Manufacturing Format (3MF)
    Used by IMF_Package, Interoperable Master Format (IMF)

Local use Explanation of format description terms

LC experience or existing holdings The Library of Congress has several workflows that rely on the ZIP format for transmitting digital content to the library. These workflows rely on open-source toolkits and/or desktop decompressors. In all current workflows, the ZIP format is used for transmission only, with the ZIP wrapper discarded after successful extraction of the contents.
LC preference To date, the Library of Congress has found no need to stipulate particular versions of ZIP_PK or particular features of the format to avoid. In general the Library of Congress does not accept content submitted in encrypted form.

Sustainability factors Explanation of format description terms

Disclosure The ZIP_PK format has been developed and maintained by PKWARE, Inc. since 1998. The format specification is proprietary, but the most recent version has always been openly disclosed as the .ZIP Application Note with a file name of APPNOTE.TXT.

In the past, PKWARE provided public access only to the most recent Application Note. In association with the standardization of Office Open XML (OOXML) as ISO 29500:2008, access was also provided to the version 6.2.0 explicitly incorporated into that format specification. PKWARE now provides access to an archive of selected versions of the .ZIP Application Note.

See Notes below for a chronology of versions of APPNOTE.TXT introducing significant new features to the format. Changes between version 6.3.2 (2007) and version 6.3.3 (2012) of APPNOTE.TXT do not reflect changes in the format, but rather editorial and layout changes based on discussions with an ISO working group that studied the ZIP format. Changes, such as adding paragraph numbers, were made to making it easier for ISO standards to reference the ZIP specification in an unambiguous and reliable way.


The ZIP_PK format has been very widely adopted. Most usage has been and continues to be based on a subset of features, resulting in a high level of interoperability. Many tools exist to create and read ZIP files that do not include strong encryption, including tools integrated into operating systems for personal computers and/or stand-alone tools shipped with new computers. See for many examples; just a few will be mentioned here.

The originator of the format, PKWARE, provides a range of products, from a free ZIP Reader for Windows, PKZIP and SecureZIP for Windows Desktop and various IBM platforms, to server-based products.

As of September 2012:

Windows ("compressed folders") and Mac OS (Archive Utility) have built in support for a subset of ZIP features. Earlier versions of these operating systems typically shipped with "lite" versions of utilities such as WinZip and Stuffit. Most versions of Unix (including the version that underlies recent Mac OS versions, ship with zip and unzip utilities, made available from Info-ZIP along with open-source software libraries compatible with many operating systems. The Info-ZIP software, developed starting in 1990 to provide ZIP-compatible tools for operating systems other than DOS, supports the most useful features supported by the ZIP_PK format, and has been widely incorporated into other products, including WinZip and JHOVE (JSTOR/Harvard Object Validation Environment). Its capabilities have thus served as a common denominator for interoperability. As of 2012, the latest version of Info-ZIP's zip utility is from 2008 and of unzip is from 2009; these versions support filenames in UTF-8 and the ZIP64 extension introduced by PKWARE in 2001 to handle large files. A Java-based utility is included in the Java Developer Kit. Apache Commons Compress extends the functionality of Several perl and python software libraries for manipulating ZIP files can be found on the web. Note: Many ZIP utilities or software libraries support only a subset of ZIP_PK features, particularly those available without licensing. Many compression/archiving utilities, including PKZIP, support ZIP_PK (.zip) as one of several input/output file formats.

Update on adoption in 2020:

ZIP support continues to extend. In 2015, ISO/IEC 21320-1 was published as an international standard for a container based on the ZIP format for digital documents that comprise multiple files, using a subset of features that can be freely implemented. See ZIP_21320_1. Support for opening ZIP files has been added to iOS. Starting with iOS 7 (2013), ZIP files received via Mail or Messages could be opened. More general support was added in 2018 with the Files app, added in iOS 11. The Files app allows users to see the contents of a ZIP file, and extract files. For Android, a number of apps exist. Comments welcome..

ZIP_PK has been used as a packaging or container format in other format specifications. These format specifications typically do not incorporate patented features of ZIP_PK and typically constrain the use of other features, for example, allowing only the DEFLATE algorithm for compression. Features that do tend to get incorporated include: long filenames, encoding of filenames, etc. in UTF-8, and the ZIP64 mechanism introduced by PKWARE for handling large ZIP files. See Notes below for more detail.

    Licensing and patents

According to Our Founder -- Phil Katz, on the PKWARE website (as of February 2013) the original developer's "decision to dedicate the .ZIP extension and file format specification to the public domain helped the .ZIP file format become a globally open standard." As a result, many utilities for creating or reading files in the ZIP_PK format have been and can be developed without separate license agreements. Two mechanisms supported by more recent versions of the format are covered by PKWARE patents: portions of the strong encryption technology support; and a technique for issuing and incorporating patches to content previously distributed as ZIP archives. Tools that manipulate ZIP files but avoid use of these mechanisms can be developed without license concerns. See Notes below for more detail on the scope of PKWARE patents and licensing.

Third parties with rights associated with code incorporated into PKWARE's own ZIP Reader products were listed, with license terms, (as of September 2012) at (link now via Internet Archive). All were made available to PKZIP under some form of public (open-source) license and would presumably be available to other tool developers. The compilers of this resource are not aware of any patent concerns with compression or encryption methods for which the ZIP_PK specification through version 6.3.3 of APPNOTE.TXT, has identifying codes. Comments welcome.

Transparency Depends upon algorithms and tools to read. Would require sophistication to build tools from scratch.
Self-documentation Since the format provides a mechanism to bundle files together, a component file that provides rich metadata in a standard or application-specific format can be included in a ZIP file. This practice has been incorporated into the package specifications for EPUB, OOXML, ODF, etc. The ZIP_PK specification includes a central directory, which contains the metadata required to extract the contents of the ZIP file, i.e., file names and directory structure. The inclusion of comments for each file is supported within the format, but not (or only partially) by some utilities or software libraries.
External dependencies None, beyond the availability of software to extract and decompress the files contained in a ZIP file. The ZIP_PK format itself includes a mechanism for defining within a ZIP file the minimum version that must be supported to extract, decompress, and decrypt (if needed) the contents. Minimum versions for particular extensions and methods for compression and encryption are documented in APPNOTE.TXT. Note: Use of strong encryption may rely on keys conveyed externally or external authentication procedures.
Technical protection considerations An important feature of the ZIP_PK format is the ability to encrypt the contained files individually. Over the years, support for increasingly strong forms of encryption, and for encryption of the ZIP central directory has been added. Certain applications that employ the ZIP format to bundle files together, such as OOXML (ISO/IEC 29500-2:2012), explicitly prohibit encryption within the ZIP-based package. Others, such as SCORM and EPUB, employ their own application-specific approaches to encryption.

Quality and functionality factors Explanation of format description terms

Compression Options for compression are extensive and include BZIP2, LZMA, DEFLATE, IBM z/OS CMPSC Compression, IBM TERSE, IBM LZ77 z Architecture, Zstandard (zstd), MP3 Compression, XZ Compression, JPEG variant, WavPack compressed data, PPMd version I, and AE-x encryption marker.
Support for Error Dectection ZIP files may use a CRC-32 hash for checking file integrity.
Beyond normal functionality Two mechanisms supported by the ZIP_PK format go beyond the functions of bundling and compression. The first is the ability to construct a self-extracting executable .zip file. The second is a "patching" mechanism, described in Notes below.

File type signifiers and format identifiers Explanation of format description terms

Tag Value Note
Filename extension zip
Other extensions are used for particular applications that use the ZIP format as a container, for example, EPUB_3, which recommends .epub as the file extension.
Internet Media Type application/zip
See registration at IANA.
File signature Hex: 50 4B 03 04
ASCII: P K etx eot
In the specification, this 4-byte "local" file header signature is expressed in little-endian form as 0x04034b50. The compilers of this resource have chosen to express the ZIP signifiers in the byte-order that is consistent with most documentation for internal file signifiers, such as PRONOM and the File Extension Source (FILext). Most ZIP files begin with this 4-byte string, since most ZIP files begin with a local file header. However, as pointed out in, a ZIP file does not need to begin with a file entry. The signature patterns below should also be considered.
File signature Hex: 50 4B 01 02 {...} 50 4B 05 06 {...}
ASCII: P K soh stx {...} P K enq ack {...}
This signature pattern adapted from PRONOM for convenient display represents the beginning of a central directory record and the mandatory end of central directory record which will be found at or near the end of a ZIP file. See PRONOM record for ZIP (PUID: x-fmt/263), which documents signatures at both the beginning and end of the file.
File signature Hex: 50 4B 05 06
ASCII: P K enq ack
At this signature based on the mandatory end of central directory record is documented as a Trailer signature, with a stipulation to expect 18 additional bytes before the end of the file. APPNOTE.TXT 6.3.3 stipulates that a valid file should have one and only one end of central directory record, but does not state explicitly that it be at the physical end of the file.
File signature Hex: 50 4B 05 06
ASCII: P K enq ack
At beginning of empty ZIP archive file. From
File signature Hex: 50 4B 07 08
ASCII: P K bel bs
At beginning of first file of a split or spanned ZIP archive. From
Pronom PUID x-fmt/263
Wikidata Title ID Q136218

Notes Explanation of format description terms


Use of ZIP file format in other specifications: The notes here describe the relationship of some specifications based on ZIP to versions and features of ZIP_PK. These notes are based on the specification documents.

  • The SCORM (Sharable Content Object Reference Model, used for learning materials) includes a Package Interchange Format (PIF) based on ZIP. The SCORM 2004 Content Aggregation Model mandates compliance with PKZip v 2.04g and RFC 1951 ( DEFLATE) as the only compression algorithm. Numbering of versions of the PKZip software and the APPNOTE.TXT specification do not correspond; however, the timelines on and suggest that SCORM 2004 usage corresponds to APPNOTE.TXT version 2.x.
  • JAR files, packages used to transmit software in the Java programming language are also based on the ZIP file format. The JAR specification was originally based on an APPNOTE.TXT version dated 1997-03-11 modified by Info-ZIP and based on a PKWARE version dated 1996-02-15. The JAR utility description also specifies that file-names must be encoded in UTF-8. Compression, if applied, uses RFC 1951 (DEFLATE). Version 7 of the Java platform (current as of September 2012) refers to the latest APPNOTE.TXT for optional use of the ZIP64 extension to support large files and for indicating the use of UTF-8 (Appendix D). The Java platform does not support encryption. [Aside: program code can be obfuscated (e.g., by changing meaningful names for variables, etc. to "a," "b,"...) to make it hard to read.]
  • The Open Packaging Conventions (OPC, Part 2 of ISO/IEC 29500, OOXML) are based on version 6.2.0 of APPNOTE.TXT. Compression is restricted to DEFLATE; digital signatures are not supported. Details on the use of ZIP in OPC are in section 10 and Annex C of ISO/IEC 29500-2:2012.
  • Version 1.2 of ODF (Open Document Format for Office Applications) from OASIS incorporates version 6.2.0 of APPNOTE.TXT. The only compression algorithm allowed is DEFLATE (RFC 1951). ODF defines its own encryption mechanism. Version 1.0 of ODF (as published as ISO/IEC 26300:2006) referred to the APPNOTE.TXT version dated 1997-03-11 from Info-ZIP, because no appropriate version was made available publicly by PKWARE at the time because the version 6.2.0 referred to by the original ODF specification from OASIS had been superseded. Now that PKWARE maintains an archive of documentation to facilitate persistent references, Version 1.2 of ODF again refers to version 6.2.0 of APPNOTE.TXT at PKWARE.
  • Versions 2 and 3 of EPUB define OCF (Open Container Format) as a package for encapsulating the components of an electronic publication. EPUB_3 mandates a container format based on ZIP as defined by version 6.3.0 of APPNOTE.TXT. In EPUB_2, the ZIP-based container was supported but not mandated. The only compression algorithm allowed is DEFLATE (RFC 1951). Split ZIP archives are explicitly prohibited. Encryption is only allowed using an EPUB-specific approach. File names, etc. must be in UTF-8. Use of the PKWARE ZIP64 extension (version 1 only) is allowed but discouraged unless the content requires it.

Bearing in mind changes in technology since the original ZIP_PK, it appears to the compilers of this description that a common set of features for the most interoperable usage is emerging, including: the basic ZIP_PK structure with CRC-32 for checking file integrity, DEFLATE compression (or no compression), and support for long filenames, for filenames, etc., encoded in UTF-8, and for large files through ZIP64. ZIP_PK features avoided in these specifications include: encryption as supported in ZIP_PK; other compression schemes; other features covered by PKWARE patents. Application-specific approaches have been developed for file manifests (with more information than in the ZIP Central Directory), metadata, digital signatures, and encryption. Comments welcome.

PKWARE patents and licensing: One set of changes between version 6.3.2 (2007) and version 6.3.3 (2012) of APPNOTE.TXT was intended to add clarity to the specification in relation to the scope and applicability of PKWARE's patents, making clearer to potential developers the particular sections of the specification that have any relationship to PKWARE's patents, with the implication that reliance on other sections should provide no patent concerns. Statements in version 6.3.3 include: "Portions of the Strong Encryption technology defined in this specification are covered under patents and pending patent applications." and "Patch support is provided by PKPatchMaker(tm) technology and is covered under U.S. Patents and Patents Pending." [See explanation of "patch technology" immediately below.] In specific sections of version 6.3.3 of APPNOTE.TXT that relate to these two features, pointers are made to Section 10 of version 6.3.3 of APPNOTE.TXT, entitled "Incorporating PKWARE Proprietary Technology into Your Product." This section includes the following statement, "The Use or Implementation in a product of APPNOTE technological components pertaining to either strong encryption or patching requires a separate, executed license agreement from PKWARE." and provides contact information. These formal statements may be read in the context of the purpose expressed in the new section 1.1.1 of version 6.3.3 of APPNOTE.TXT, "This specification is intended to define a cross-platform, interoperable file storage and transfer format. Since its first publication in 1989, PKWARE, Inc. ("PKWARE") has remained committed to ensuring the interoperability of the .ZIP file format through periodic publication and maintenance of this specification. We trust that all .ZIP compatible vendors and application developers that use and benefit from this format will share and support this commitment to interoperability."

PKWARE declares the patents it owns at

Patching technology support in ZIP format: Jim Peterson of PKWARE provided the following explanation: The "patching" technology is a means for using ZIP files to distribute revised document content by delivering only the changed elements of a prior document instead of having to deliver a complete new copy of the revised version. The premise is that a user delivers version 1 of a file to other users. Later, version 2 is prepared for delivery to the same recipients. The patching technology provides capabilities to compare version 1 and version 2 and distill out only the differences between the documents. The user distributing version 2 can then deliver only the differences rather than a whole new document with the objective of reducing the amount of information to deliver to update a user from version 1 to version 2. Upon receipt, the changes can be applied directly the original version 1 of the document/file with the end result that version 2 becomes available to the receiving users through the patching technology.

Support for ZIP_PK features: The 2012 version 6.3.3 of the .ZIP Application Note indicates, "Newer versions of this format specification are intended to remain interoperable with all prior versions whenever technically possible." Although PKWARE products are likely to support all features in ZIP_PK, other ZIP utilities will not necessarily support optional features. In particular, support for any particular compression or encryption method is optional. Section 4 of the 2012 version 6.3.3 of APPNOTE.TXT provides a clearer specification of which features are mandatory and which optional than earlier versions. Section 4.1.3 states, "Compression method 8 (Deflate) is the method used by default by most Zip compatible application programs." Apache Commons Compress, a software toolkit built in Java, provides some informative details on creating interoperable ZIP files at

Variants from other vendors: Due to differences in timing of introduction of features in different products, such as WinZip and the Info-ZIP tools, some files with a .zip extension produced in the past may not strictly comply with the ZIP_PK specification as represented by a particular edition of APPNOTE.TXT. Others may use the mechanisms for extensibility in ZIP_PK in ways that other utilities do not recognize. In general, a consensus towards convergence has emerged in recent years. For example, features introduced by other developers have been folded back into the ZIP_PK specification. The compilers of this resource are unable to determine how serious the potential problem of incompatible legacy files is in practice. Experience at the Library of Congress with the use of ZIP files for transmission of digital content to the library has been almost entirely trouble-free.

The WinZip product introduced a .zipx extension to distinguish archive files that used more advanced features, retaining .zip for files using only the features supported by most ZIP utilities. Version 6.3.3 of APPNOTE.TXT mentions .zipx as an extension that may be used for files compliant with the format it specifies. The compilers of this resource are not aware of any use of the .zipx extension for files created by tools other than WinZip. Comments welcome.

Several different tools that have "zip" in their names have their own native storage format and do not necessarily support the ZIP_PK format described here, for example, gzip; 7-zip; and bzip2.

Data integrity checks: A mandatory feature of the ZIP_PK format is the use of a 32-bit cyclic redundancy check (CRC) for each file using a specific CRC-32 algorithm. On decompression, ZIP utilities are expected to recalculate a CRC on the decompressed file and check against the CRC stored in the ZIP archive file. In addition, ZIP_PK provides a way to store optional digital signatures.


Chronology for APPNOTE.TXT versions: Since the first publication of APPNOTE.TXT in 1990, ZIP_PK has been extended in response to a changing technological environment, while retaining its straightforward structure. As new features have been added to ZIP_PK, care has been taken to ensure backwards compatibility, so that new software versions can read files in all previous versions of the format specification. A chronology of significant versions of the APPNOTE specification for ZIP_PK, prior to and including 2012, follows below. The details and descriptions were provided by Jim Peterson, Chief Scientist, PKWARE. Versions not on this list introduced only minor changes, such as corrections or support for an additional compression method.

  • 1.0, 1990-03-15 []
    Initial ZIP format created by Phil Katz of PKWARE, Inc. Includes original 96-bit (triple 32-bit) password encryption storage definition.
  • 2.0, 1993-02-01 []
    Introduces the Deflate data compression algorithm which becomes one of the most widely used compression algorithms.
  • 4.0, 2000-11-01 []
    Introduces support for using digital signatures to verify data in a ZIP file and adds the fast Deflate64 data compression algorithm.
  • 4.5, 2001-11-01 []
    Introduces support for storing more than 65535 files in a ZIP file and file sizes exceeding the previous 4 gigabyte limit.
  • 5.2, 2003-07-16 []
    Introduces initial storage specification for adding stronger encryption using algorithms exceeding 128-bits based on passwords, digital certificates, or a combination of both simultaneously.
  • 6.2.0, 2004-04-26 []
    Defines storage specification to support encrypting ZIP file metadata, such as file names, within the ZIP Central Directory. Reference version for incorporating the ZIP format into Office Open XML File Format standard ECMA-376 (versions 1, 2, and 3) and ISO/IEC 29500 (versions published in 2008 and 2012).
  • 6.2.2, 2006-01-06 []
    Documents final storage specification for adding stronger encryption using algorithms exceeding 128-bits using passwords, digital certificates, or a combination of both simultaneously.
  • 6.3.0, 2006-09-29 []
    Documents support within ZIP files for using hash algorithms of SHA-256, SHA-384 and SHA-512; incorporates LZMA and PPMd data compression algorithms; support for Blowfish and Twofish data encryption algorithms; definition of UTF-8 international file name storage.
  • 6.3.3, 2012-09-01 []
    Formatting changes to the document add conformance requirements and support easier referencing by other standards that incorporate the ZIP format.

ZIP_PK in the international standards environment: As described in, a proposed project to create an ISO/IEC international standard for a format compatible with ZIP failed to pass a 2010 ballot of national standards bodies. Instead, a study period was initiated, resulting in recommendations documented in ISO/IEC JTC 1/SC 34 N 1621. The recommendations were (a) to have PKWARE continue its maintenance of the ZIP Application Note, (b) to plan for a new multi-part ISO standard to build on top of the ZIP Application Note, and (c) to propose a new work item for Part 1 of the new standard for a Document Container File. The new work item was approved in August 2011 and the standard was published in 2015, as ISO/IEC 21320-1:2015 - Information technology -- Document Container File -- Part 1: Core.

Format specifications Explanation of format description terms

Useful references


Last Updated: 01/02/2024