First principles ... XML.

What is a Document?

Outside the IT world, we encounter all sorts of hard-copy documents: letters, forms, books, newspapers, magazines, invoices, maps, birthday cards, leaflets, posters, sticky notes, and many others. The concept of the hard-copy document evolved almost without notice. When we encounter new types of hard- copy documents in our homes, offices, classrooms, libraries, stationery stores, or local newsstands, we seem to accept them unconsciously. Meanwhile, within the IT world, the concept of the electronic document has evolved, too. Let's start with a more basic definition first:

The definition of text. Text is generally considered to consist of words, sentences, lines, paragraphs, and even pages. Typically, the term text also refers to electronic text stored as only simple character codes (for example, American Standard Code for Information Interchange, or ASCII, codes)-that is, without any formatting. At one time, the electronic document was considered to be only a text file created with applications called text editors or word processors. You could almost use the terms text and document interchangeably. However, as developments occurred on many IT fronts, the concept of the electronic document expanded to contain tables, graphics, charts, and other objects, in a manner that parallels the evolution of hard-copy documents. Now, in the IT world, documents are considered to be electronic files of any size for any media (for example, text, audio, video, and graphics), created by any application. Thus now, the definition of text is a subset of the definition of the document.

In their Extensible Markup Language 1.0 Recommendation, which is recognized as the official XML standard, the W3C (World Wide Web Consortium) defines an XML document as a 'data object if it is well-formed, as defined in (Extensible Markup Language Recommendation).... Each XML document has both a logical and a physical structure.'

Related to the discussion of documents is the term document processing, which is the discipline that deals with creating applications that allow you to deal with documents of all types. Document processing is split into creating or manipulating those documents destined for human viewing and consumption (people-oriented processing), as well as those that are destined for computer consumption (machine-oriented processing). Documents of the former type were comparatively long-lived (examples: specifications, drawings, procedures, charts, and memos). Documents of the latter type tend to have shorter lives because their data may be manipulated, transformed, or combined on the fly to create or add to different documents. XML descends from a rich document-processing heritage.

What Is Markup?

The concept of markup is important. After all, it's the M in XML. But what does it mean? Basically, it's a way to add information about data to the data itself. You may not have had much experience with other markup languages, but you have probably used markup in one form or another. For example, have you ever:

* Underlined or highlighted words or passages on a hard-copy document to indicate important information?

* Marked up a draft hard copy of a document with symbols indicating "new paragraph here," "bold this,' or 'remove this"?

* Made marks on a map indicating where you want to turn, or where specific features are located?

* Numbered bits of information, such as steps, in an otherwise unnumbered procedure?

Those and similar activities involve marking up data. All the symbols, notes, numbers, designated actions, or highlights-all of which qualify as some sort of markup--emphasize or convey something about the data: what it means or what you are supposed to do with it.

A significant paper entitled 'Markup Systems and the Future of Scholarly Text Processing," by James H. Coombs and Allen H. Renear of Brown University and Steven J. DeRose of Electronic Book Technologies, describes six types of markup:

* Punctuational, which consists of the use of defined marks (examples: spaces, periods, and commas) to provide primarily syntactic information about written utterances. Punctuation has been around so long that we take it for granted.

* Presentational, which we use to group our materials for order and clarity. Examples include horizontal and vertical spacing, page breaks, numbering, chapter and section breaks, justification, and lists.

* Procedural, which is a characteristic of whatever system will be used to create presentations. Often grouped with what we call file formats, it tells someone or something (such as a formatter with a set of installed drivers) about the size and format of a document (examples: letter, legal, and portrait and landscape views), fonts, and other production information.

* Descriptive, which allows, authors to identify certain elements of their data as belonging to a specific family of text. The common word- processing tag BT (for basal text) is au example: When a text formatter encounters that code, it consults, and then follows, a predefined set of roles that tell it what to do to display or print the characters associated with that code. If changes become necessary, you only need to change the roles, not each BT tag in the document.

* Referential, which refers to separate physical or electronic entities (that is, located external to the document being processed) that will be imported and placed in the proper sequence during document processing.

* Metamarkup, which provides the ability to control the definition and interpretation of markup tags, and to extend the vocabulary of derivative markup languages. Metadata, the concept of information about information, is related to this concept, if you would like to read the Coombs, Renear, and DeRose paper that the preceding definitions were taken from, you can find it online at www.oasis-open.org/cover/coombs.html#Figure1.

Markup, in summary, is the inserting of characters or symbols into a document to indicate the document's physical and logical structure, to indicate how the information in a document should appear, or to provide some other form of instruction. The primary goal of markup is to separate the treatment (for example, the appearance or structure) of a document from the actual data in the document.

XML Is a Markup Language and a Metalanguage 

There are over two dozen categories of computer languages; you are probably familiar with some of them already. For example, machine languages consist entirely of numbers and are only understood by computers; assembly languages are symbolic representations of the machine language of a specific computer; programming languages such as COBOL, C++, Java, and Fortran instruct computers to do specific tasks; and fourth-generation languages, whose syntax is closer to human languages.

Some language categories are separate and discrete, dedicated to specific functions; some languages are subsets of others; and some are hybrids of other languages.

For a more comprehensive listing of computer languages and their respective definitions, consult the The Language List Web site, maintained by Bill Kinnersley of the Computer Science Department, University of Kansas, at http:llcui.unige.ch/OSG/info/Langlistlintro.htmi.

XML doesn't fall into any of the categories previously listed, but it falls into two other categories: It's a markup language and a metalanguage.

Markup Languages

Extrapolating the definition of markup, markup languages are those that allow us to create documents consisting of plaint ext data and other entities, plus markup codes that define the logical components and structure, as well as describe the appearance or other aspects of the data. The markup codes, called tags, are located adjacent to their respective data. In addition, the data and tags are usually composed of common text characters, so they can remain independent of platform and operating system. Why use markup languages? These days, with the proliferation of computer networks across the world, with their myriad of applications, operating systems, and proprietary network devices, the data transmitted over the wire, through the air, and through space must include all the information necessary for automated systems (such as computers, routers, firewalls, and hubs) to transmit, receive, and otherwise deal with the data. The receiver needs the markup tags to interpret the message: the format and content of database data, multimedia graphic files or audio files, debit card transactions, credit card authorizations, or any other various document types.

Metalanguages

In the What Is Markup? section, we provided a listing of markup types. One of the types was called metamarkup, which provides the capability to control the definition and interpretation of markup tags, and to extend the vocabulary of derivative markup languages. That is consistent with the definition found at Mr. Kinnersley's Web site, where he defines a metalanguage as a 'language used for formal description of another language." It is also consistent with other definitions of metalanguages, which describe them as languages that provide for conformance-proving mechanisms. XML permits developers to create their own specialized derivative languages, but all of those languages have one thing in common: They meet XML specifications. If languages and documents contravene the XML specifications, the XML processors in their respective applications may or may not process them. Even if they do, they will likely generate error messages.

From XML in 60min per day. Wiley ISBN 0-4711-42254-1