PDF: Difference between revisions

'''Portable Document Format''' ('''PDF'''), standardized as '''ISO 32000''', is a [[file format]] developed by [[Adobe Inc.|Adobe]] in 1992 to present [[document]]s, including text formatting and images, in a manner independent of [[application software]], [[Computer hardware|hardware]], and [[operating system]]s.<ref name="pdf-ref-1.7">{{cite web|author=Adobe Systems Incorporated|url=https://www.adobe.com/devnet/acrobat/pdfs/pdf_reference_1-7.pdf|title=PDF Reference|date=November 2006|edition=6th|version=1.7|url-status=dead|archiveurl=https://web.archive.org/web/20081001170454/https://www.adobe.com/devnet/acrobat/pdfs/pdf_reference_1-7.pdf|archivedate=October 1, 2008|accessdate=January 12, 2023}}</ref><ref>{{Cite web|last=Warnock|first=J.|url=https://www.pdfa.org/norm-refs/warnock_camelot.pdf|title=The Camelot Project|date=14 October 2004<!--dates from PDF source-->|orig-date=Original date 5 May 1995|archive-url=https://web.archive.org/web/20110718230852/http://www.planetpdf.com/planetpdf/pdfs/warnock_camelot.pdf|archive-date=July 18, 2011|url-status=live}}</ref> Based on the [[PostScript]] language, each PDF file encapsulates a complete description of a fixed-layout flat document, including the text, [[font]]s, [[vector graphics]], [[raster images]] and other information needed to display it. PDF has its roots in "The Camelot Project" initiated by Adobe co-founder [[John Warnock]] in 1991.<ref>{{Cite web|title=What is a PDF? Portable Document Format {{!}} Adobe Acrobat DC|url=https://www.adobe.com/acrobat/about-adobe-pdf.html|access-date=January 12, 2023|publisher=Adobe Systems Inc.|language=en|archive-date=January 30, 2023|archive-url=https://web.archive.org/web/20230130032548/https://www.adobe.com/acrobat/about-adobe-pdf.html|url-status=live}}</ref>

In April 2023 the PDF Association made ISO 32000-2 available for download free of charge.<ref name=nowfree>{{cite press release| title=Announcing no-cost access to the latest PDF standard: ISO 32000-2 (PDF 2.0)| publisher=PDF Association| url=https://pdfa.org/sponsored-standards | date=16 June 2023| orig-date=Updated; originally published 5 April 2023| access-date=October 6, 2023| archive-date=September 23, 2023| archive-url=https://web.archive.org/web/20230923202322/https://pdfa.org/sponsored-standards/| url-status=live}}</ref>

== {{Anchor|Technical foundations}}Technical details ==

[[PostScript]] is a [[page description language]] run in an [[Interpreter (computing)|interpreter]] to generate an image.<ref name="Pfiffner_Page_137" /> It can handle graphics and has standard features of [[programming language]]s such as [[conditional (computer programming)|branching]] and [[loop (computing)|looping]].<ref name="Pfiffner_Page_137" /> PDF is a subset of PostScript, simplified to remove such [[control flow]] features, while graphics commands remain.<ref name="Pfiffner_Page_137" />

PostScript was originally designed for a drastically different [[use case]]: transmission of one-way linear [[print job]]sjobs in which the PostScript interpreter would collect a series of commands until it encountered the <code>showpage</code> command, then execute all the commands to render a page as a raster image to a printing device.<ref name="Pfiffner_Page_139">{{cite book |last1=Pfiffner |first1=Pamela |title=Inside the Publishing Revolution: The Adobe Story |date=2003 |publisher=Peachpit Press |location=Berkeley |isbn=0-321-11564-3 |page=139}}</ref> PostScript was not intended for long-term storage and real-time interactive rendering of [[electronic document]]s to [[computer monitor]]s, so there was no need to support scrollinganything backother tothan previousconsecutive rendering of pages.<ref name="Pfiffner_Page_139" /> If there was an error in the final printed output, the user would correct it at the application level and send a new print job in the form of an entirely new PostScript file. Thus, any given page in a PostScript file could be accurately rendered only as the cumulative result of executing all preceding commands to draw all previous pages—any of which could affect subsequent pages—plus the commands to draw that particular page, and there was no easy way to bypass that process to skip around to different pages.<ref name="Pfiffner_Page_139" />

Traditionally, to go from PostScript to PDF, a source PostScript file (that is, an executable program) is used as the basis for generating PostScript-like PDF code (see, e.g., [[Adobe Distiller]]). This is done by applying standard [[compiler]] techniques like [[loop unrolling]], [[inline expansion|inlining]] and removing unused branches, resulting in code that is purely declarative and static.<ref name="Pfiffner_Page_139" /> The end result is then packaged into a [[container format]], together with all necessary [[Dependency (computer science)|dependencies]] for correct rendering (external files, graphics, or fonts to which the document refers), and [[Data compression|compressed]]. Modern applications write to printer drivers which directly generate PDF rather than going through PostScript first.

* PDF contains only static [[Declarative programming|declarative]] PostScript code that can be processed as data, and does not require a full program [[Interpreter (computing)|interpreter]] or [[compiler]].<ref name="Pfiffner_Page_139" /> This avoids the complexity and security risks of an engine with such a higher complexity level.

* PDF enforces the rule that the code for any particular page cannot affect any other pages.<ref name="Pfiffner_Page_139" /> That rule is strongly recommended for PostScript code too, but has to be implemented explicitly (see, e.g., the [[Document Structuring Conventions]]), as PostScript is a full programming language that allows for such greater flexibilities and is not limited to the concepts of pages and documents.

* LossA loss of flexibility, and limitation to a single use case.{{cncitation needed|date=December 2023}}

An index table, also called the cross-reference table, is located near the end of the file and gives the byte offset of each indirect object from the start of the file.<ref>Adobe Systems, PDF Reference, pp. 39–40.</ref> This design allows for efficient [[random access]] to the objects in the file, and also allows for small changes to be made without rewriting the entire file (''incremental update''). Before PDF version 1.5, the table would always be in a special ASCII format, be marked with the <code>xref</code> keyword, and follow the main body composed of indirect objects. Version 1.5 introduced optional ''cross-reference streams'', which have the form of a standard stream object, possibly with filters applied. Such a stream may be used instead of the ASCII cross-reference table and contains the offsets and other information in binary format. The format is flexible in that it allows for integer width specification (using the <code>/W</code> array), so that for example, a document not exceeding 64&nbsp;[[KiB]] in size may dedicate only 2&nbsp; bytes for object offsets.

Page dimensions are not limited by the format itself. However, Adobe Acrobat imposes a limit of 15 million by 15 million inches, or 225 trillion in² (145,161 &nbsp;km²).<ref name="pdf-ref-1.7" />{{rp|1129}}

[[Raster graphics|Raster images]] in PDF (called ''Image XObjects'') are represented by dictionaries with an associated stream. The dictionary describes the properties of the image, and the stream contains the image data. (Less commonly, small raster images may be embedded directly in a page description as an ''inline image''.) Images are typically ''filtered'' for compression purposes. Image filters supported in PDF include the following general-purpose filters:

=== Logical structure and accessibility<span class="anchor" id="Tagged PDF"></span> ===

A "'''tagged" PDF'''<!--boldface per [[WP:R#PLA]]--> (see clause 14.8 in ISO 32000) includes document structure and semantics information to enable reliable text extraction and [[accessibility]].<ref>{{cite web |title=Tagged PDF Best Practice Guide: Syntax |url=https://pdfa.org/wp-content/uploads/2019/06/TaggedPDFBestPracticeGuideSyntax.pdf |website=pdfa.org |publisher=[[PDF Association]] |date=June 2019 |access-date=2024-06-24}}</ref> Technically speaking, tagged PDF is a stylized use of the format that builds on the logical structure framework introduced in PDF 1.3. Tagged PDF defines a set of standard structure types and attributes that allow page content (text, graphics, and images) to be extracted and reused for other purposes.<ref>{{cite web|first=Duff|last=Johnson|date=April 22, 2004|title=What is Tagged PDF?|url=https://www.talkingpdf.org/what-is-tagged-pdf/|url-status=live|archive-url=https://web.archive.org/web/20040807132851/http://www.planetpdf.com/enterprise/article.asp?ContentID=6067|archive-date=August 7, 2004}}</ref>

Tagged PDF is not required in situations where a PDF file is intended only for print. Since the feature is optional, and since the rules for Taggedtagged PDF were relatively vague in ISO 32000-1, support for tagged PDF among consuming devices, including [[assistive technology]] (AT), is uneven as of 2021.<ref>{{Cite web|title=Is PDF accessible?|website=DO-IT - Disabilities, Opportunities, Internetworking, and Technology|publisher=University of Washington|date=October 4, 2022|url=https://www.washington.edu/doit/pdf-accessible?1002=|access-date=January 12, 2023|archive-date=February 10, 2023|archive-url=https://web.archive.org/web/20230210114239/https://www.washington.edu/doit/pdf-accessible?1002=|url-status=live}}</ref> ISO 32000-2, however, includes an improved discussion of tagged PDF which is anticipated to facilitate further adoption.

PDF files can be infected with viruses, Trojans, and other malware. They can have hidden JavaScript code that might exploit vulnerabilities in a PDF, hidden objects executed when the file that hides them is opened, and, less commonly, a malicious PDF can launch malware.<ref>{{cite web | title=Can PDFs have viruses? Keep your files safe | publisher=Adobe | url=https://www.adobe.com/acrobat/resources/can-pdfs-contain-viruses.html | access-date=3 October 2023 | archive-date=October 4, 2023 | archive-url=https://web.archive.org/web/20231004120143/https://www.adobe.com/acrobat/resources/can-pdfs-contain-viruses.html | url-status=live }}</ref>

There are many software options for creating PDFs, including the PDF printing capabilities built into [[macOS]], [[iOS]],<ref>{{Cite web|url=https://ijunkie.com/how-to-create-pdf-web-page-safari-iphone-ipad-ios-11/|title=How to Create a PDF from Web Page on iPhone and iPad in iOS 11|last=Pathak|first=Khamosh|date=October 7, 2017|website=iJunkie|access-date=January 12, 2023|archive-date=January 12, 2023|archive-url=https://web.archive.org/web/20230112153246/https://ijunkie.com/how-to-create-pdf-web-page-safari-iphone-ipad-ios-11/|url-status=live}}</ref> and most [[Linux]] distributions. Much document processing software including [[LibreOffice]], [[Microsoft Office 2007]] (if updated to [[Office 2007#Service Pack 2|SP2]]) and later,<ref>{{cite web|url=http://support.microsoft.com/kb/953195|title=Description of 2007 Microsoft Office Suite Service Pack 2 (SP2)|publisher=[[Microsoft]]|url-status=dead|archive-url=https://web.archive.org/web/20090429212434/http://support.microsoft.com/kb/953195|archive-date=April 29, 2009|access-date=January 12, 2023}}</ref> [[WordPerfect]] 9, and [[Scribus]] can export documents in PDF format. There are many PDF print drivers for Microsoft Windows, the [[pdfTeX]] typesetting system, the [[DocBook]] PDF tools, applications developed around [[Ghostscript]] and [[Adobe Acrobat]] itself as well as [[Adobe InDesign]], [[Adobe FrameMaker]], Adobe Illustrator, Adobe Photoshop, that allow a "PDF printer" to be set up, which when selected sends output to a PDF file instead of a physical printer. [[Google]]'s online office suite [[Google Docs]] allows uploading and saving to PDF. Some web apps offer free PDF editing and annotation tools.

The [[Free Software Foundation]] werewas "developing a free, high-quality and fully functional set of libraries and programs that implement the PDF file format and associated technologies to the ISO 32000 standard", as one of theirits [[High priority free software projects|high priority projects]].<ref>On 2014-04-02, a note dated February 10, 2009 referred to [http://www.fsf.org/campaigns/priority.html Current FSF High Priority Free Software Projects] {{Webarchive|url=https://web.archive.org/web/20070810230457/http://www.fsf.org/campaigns/priority.html |date=August 10, 2007 }} as a source. Content of the latter page, however, changes over time.</ref><ref>{{cite web|url=http://gnupdf.org/Goals_and_Motivations|title=Goals and Motivations|publisher=GNUpdf|date=November 28, 2007|website=gnupdf.org|access-date=April 2, 2014|archive-date=July 4, 2014|archive-url=https://web.archive.org/web/20140704114405/http://www.gnupdf.org/Goals_and_Motivations|url-status=live}}</ref> In 2011, however, the GNU PDF project was removed from the list of "high priority projects" due to the maturation of the [[Poppler (software)|Poppler library]],<ref>{{cite web|title=GNU PDF project leaves FSF High Priority Projects list; mission complete!|url=http://www.fsf.org/blogs/community/gnu-pdf-project-leaves-high-priority-projects-list-mission-complete|date=October 6, 2011|first=Matt|last=Lee|publisher=Free Software Foundation|website=fsf.org|archive-date=December 28, 2014|archive-url=https://web.archive.org/web/20141228050435/http://www.fsf.org/blogs/community/gnu-pdf-project-leaves-high-priority-projects-list-mission-complete|url-status=live}}</ref> which has enjoyed wider use in applications such as [[Evince]] with the [[GNOME]] desktop environment. Poppler is based on [[Xpdf]]<ref>{{cite web|url=http://poppler.freedesktop.org/|title=Poppler Homepage|quote=Poppler is a PDF rendering library based on the xpdf-3.0 code base.|access-date=January 12, 2023|archive-date=January 8, 2015|archive-url=https://web.archive.org/web/20150108235708/http://poppler.freedesktop.org/|url-status=live}}</ref><ref>{{cite web|url=http://cgit.freedesktop.org/poppler/poppler/tree/README-XPDF|title=Xpdf License|quote=Xpdf is licensed under the GNU General Public License (GPL), version 2 or 3.|access-date=January 12, 2023|archive-date=April 14, 2013|archive-url=https://archive.today/20130414194348/http://cgit.freedesktop.org/poppler/poppler/tree/README-XPDF|url-status=live}}</ref> code base. There are also commercial development libraries available as listed in [[List of PDF software]].

In 1993, the Jaws raster image processor from Global Graphics became the first shipping prepress RIP that interpreted PDF natively without conversion to another format. The company released an upgrade to theirits Harlequin RIP with the same capability in 1997.<ref>{{cite web |url = http://www.globalgraphics.com/products/harlequin-multi-rip |title=Harlequin MultiRIP|access-date=March 2, 2014|url-status=dead|archive-url=https://web.archive.org/web/20140209215413/http://www.globalgraphics.com/products/harlequin-multi-rip/|archive-date=February 9, 2014 }}</ref>

=== Conversion and Information Extraction ===

PDF's emphasis on preserving the visual appearance of documents across different software and hardware platforms poses challenges to the conversion of PDF documents to other [[Filefile format | file formats]]s and the targeted [[Information extraction | extraction of information]], such as text, images, tables, [[Bibliography | bibliographic information]], and document [[Metadata | metadata]]. Numerous tools and source code libraries support these tasks. Several labeled [[dataset | datasets]]s to test PDF conversion and information extraction tools exist and have been used for benchmark evaluations of the tool's performance.<ref>{{Citation |last=Meuschke |first=Norman |title=A Benchmark of PDF Information Extraction Tools Using a Multi-task and Multi-domain Evaluation Framework for Academic Documents |date=2023 |work=Information for a Better World: Normality, Virtuality, Physicality, Inclusivity |volume=13972 |pages=383–405 |editor-last=Sserwanga |editor-first=Isaac |url=https://link.springer.com/10.1007/978-3-031-28032-0_31 |place=Cham |publisher=Springer Nature Switzerland |language=en |doi=10.1007/978-3-031-28032-0_31 |isbn=978-3-031-28031-3 |last2=Jagdale |first2=Apurva |last3=Spinde |first3=Timo |last4=Mitrović |first4=Jelena |last5=Gipp |first5=Bela |editor2-last=Goulding |editor2-first=Anne |editor3-last=Moulaison-Sandy |editor3-first=Heather |editor4-last=Du |editor4-first=Jia Tina|arxiv=2303.09957 }}</ref>

* {{cite book | last1 = Hardy | first1 = M. R. B. | last2 = Brailsford | first2 = D. F. | chapter = Mapping and displaying structural transformations between XML and PDF | title = Proceedings of the 2002 ACM symposium on Document engineering – DocEng '02 | pages = 95–102 | year = 2002 | url = https://www.cs.nott.ac.uk/~psadb1/Publications/Download/2002/Hardy02.pdf | doi = 10.1145/585058.585077 | publisher = Proceedings of the 2002 ACM symposium on Document engineering |isbn = 1-58113-594-7 | s2cid = 9371237 }}{{relevant?relevance inline|date=May 2022|reason=Why would random conference paper about some particular plugin for Adobe Acrobat be of interest to the reader?}}