Typographically speaking, the difference between calligraphy and script is that script has letterforms which look connected or with a suggestion that they were meant to be, whereas calligraphy has unconnected letterforms or letterforms which look like they were intended to be unconnected.

Some typeface designs are calligraphic in appearence. These include: American Uncial, Banco, Dom, Lydian, Ondine, and Reporter No. 2.

There are five classification of case: UPPERCASE, Titlecase, lowercase, mixed, and caseless. UPPERCASE and lowercase have commonly understood meanings, but Titlecase is uncommon and arises with digraphs, trigraphs, or n-graphs as follows. Considering a digraph 'DZ' (which occurs in Eastern European languages), the UPPERCASE would be 'DZ', the lowercase would be 'dz', and the Titlecase, used where the first letterform only is to be capitalized, would be 'Dz'. Since a digraph, trigraph, or n-graph is managed as a single character, all three forms are distinct, and should be represented in a font resource. The mixed classification applies to alphabets which are comprised of a mixture of UPPERCASE and lowercase letterforms, each letter only having one case form rather than two. An example of this is a Single Alphabet Uncial typeface. The caseless classification applies to letterforms which have only a single case, but which are incorporated in alphabets or scripts which have the property of case. Examples would be letterforms comprising the IPA phonetic alphabet or certain letterforms used in African and Turkish variants of the Roman alphabet which only have a single case classification form.

Character glyphs which are usually considered caseless include such symbols as mathematical operators, monetary units, numerals, and punctuation. [These symbols, however, are sometimes designed with UPPERCASE and lowercase ( « small cap » ) styles so as to appear more size consistent in text using UPPERCASE and lowercase character glyphs; but they are still functionally caseless.]

Text strings which are considered case sensitive undergo character changes when the case is changed. A character change could be a one-to-one, one-to-many, or many-to-one.

Numerals normally are caseless. However, in today's modern electronic information age, numerals are often mixed with alphabet letterforms. Since software often treats a text string or text character depending on its case, numerals sometimes need to be assigned a case class depending on the circumstances. For example, in the case of text used in certain computer language instructions, numerals might be considered caseless; whereas for text in a database which may be searched, numerals might be considered as UPPERCASE.

Examples of the use of case folding include (but are not limited to):
(1) Search processes
If the search text string has UPPERCASE only or both UPPERCASE and lowercase characters, a lowercase character is substituted for each UPPERCASE character in the string before the string is submitted for search. This is termed case conversion, and is the most basic strategy for implementing case folding.
(2) Electronic publishing document layout
If a heading text string, typically having been set with UPPERCASE beginning word characters and lowercase for the remaining characters, is to be reset as all UPPERCASE (where UPPERCASE characters are substituted for all lowercase characters) or the heading text string is to be reset as « all caps » (where « small cap » characters replace all lowercase characters).
(3) Internationalized Internet domain naming
A domain name string is to be represented only in a single case, achievable through case folding. In 2000, the protocall or model to achieve this has yet to be determined by the international standards organizations involved.

The process of substituting a lowercase character for an UPPERCASE/Titlecase character may not be straightfoward for the following reasons:
(1) An UPPERCASE character may not have a lowercase equivalent in a particular encoding. [An example of this is the old Greenlandic character 'kra' (at Unicode <0138>)].
(2) Different languages which share the same script may not have the same UPPERCASE/lowercase equivalent pairs. [An example of this is French, which, by convention, uses both E/é or É/é, but English would always use É/é in foreign-derived words.]
(3) Some languages might use different pairs. [An example of this is Turkish, which uses 'Idot'/i and I/'dotlessi' as pairs, but English uses I/i as a pair.]
(4) Some script characters may not have a lowercase form or they may be caseless.

The process of substituting an UPPERCASE character for a lowercase character may not be straightfoward for the following reasons:
(1) An UPPERCASE character may not exist in a particular character set or encoding. [An example is 'Ydieresis'/ÿ, where ÿ exists in the ISO 8859-1 character set, but 'Ydieresis' does not exist.]
(2) An UPPERCASE ligature character equivalent to a lowercase ligature character may not exist, and therefore a one-to-one character substitution cannot be made. [Examples of this type are fi or ß. These must be replaced by UPPERCASE character pairs FI and SS, respectively.]

An advantage of using CGI scripts is that the actual script commands are inaccessible to casual viewing or copying, since they are not downloaded in order to be executed. This advantage, however, is only due to one of the many major failures of the W3C committee in building needed capabilites (for file security, in this case) into HTML.

It was originally developed by Ares Software Corp. to enable an end-user to generate variations on the design of a font, and then enable more fonts to be used with a given amount of printer memory. Adobe Systems later purchased Ares Software, and has built Chameleon into PostScript Level 3. There, it is used to compress font data for the standard core fonts usually supplied with a PostScript interpreter. Adobe previously developed a technology (MultipleMaster format) which enables an end-user to alter or change certain design aspects of a font. Several foundries, including Production First Software, produce fonts in the MultipleMaster format.

The disadvantage of Chameleon technology is that it is « lossy » and therefore cannot render outlines with arbitrarily high accuracy, introducing artifacts. These artifacts are in addition to artifacts generated by the digitizing error inherent in output devices.

Formally, the object correctly designated by the term « character » is independent of style and is a designated number, which then is used as a pointer or index to the specific object (a set of software instructions or data, a cast metal impression block, etc.) which is the actual agent for making visible marks. The term is often applied colloquially and erroneously to the visible shape (or glyph) which is the result of the display of a character by use of an agent, such as a font or impression block.

Characters have various properties such as: canonical ordering; case; combining class; directionality; letter class; mathematical; mirroring; name; numeric value class; spacing; special; and width. The name property is used internal to fonts resources and other software in the process of rendering text. The numeric value is the way characters are specified in encoding vectors and, oftentimes, text.

An example of a one-to-one mapping would be the text character [65]_decimal or <0041> which would map to a glyph of the Latin letter A.

An example of a one-to-many would be the precomposed text character [225]_decimal or <00e1> mapping to glyphs a and ' so as to render as á. This could also be handled by using two text characters sequentially (mapping to a and ' in a one-to-one); but there are typographic problems with this approach.

An example of a many-to-one mapping would be two text characters [102]_decimal or <0066> (which maps to the glyph f alone), and [108]_decimal or <006c> (which maps to the glyph l alone), mapping to the ligature glyph fl.

If the example above for the one-to-many mapping were recast to illustrate a mapping which would render í, then an additional complication would be revealed: the character [105]_decimal or <0069> cannot be used to map to the i glyph as the base, because the base ' i ' glyph must not have a dot over it. Therefore, the glyph actually used must be 'dotlessi', or an ' i ' without a dot. On the otherhand, a composite character using a diacritical mark below the baseline still requires that the base character glyph i have the dot over it. Since both 'i' and 'dotlessi' are present in Unicode, this is not a distinction which must be made based on a base glyph variant, but rather a different base character. This requires an additional mapping, although this case can still be represented by algorithmic mapping rules. In the case of the one-to-many mapping of  ' j ' with ' circumflex ', the situation gets even more complicated because a 'dotlessj' is required as the base glyph, and this is not encoded in Unicode.

Sometimes for a composite character, the base character glyph also changes depending on the shape of the diacritical mark. This is a discretionary typeface style-dependent trait, thereby not able to be recognized by an algorithmic mapping rule, because base glyph variants must be used. An example of this can be found in the Production First Software typeface LafayettePF. But this must be handled by the rendering engine (or, hypothetically, by an extremely smart font resource) or by some other means than by using the mechanisms identified by the character/glyph model.

One of the fundamental principles upon which the character/glyph model is based is that an encoding should be used to represent characters spelled in text and used as lexical elements, with no notion of mapping glyph variants, ligatures, or uncommon dingbats. This principle, however, has not been strictly adhered to, and, in fact, has been broken often in many encodings (even in the newer encodings such as Unicode), because of the reality that application software which can perform glyph mapping management has not been widely available. An alternative to the character/glyph model is the direct encoding model.

The principles of the character/glyph model were first enabled with the development of PostScript, and its font machinery. PostScript fonts use the character/glyph model with the notion of encoding and the capability of reëncoding. Applications which can perform automatic ligature substitution would also make use of this model.

(another definition follows)

(another definition follows)

The CJK character collections can be viewed as a « dense » array with « sparse » access (because few consecutively-used ideographs are located adjacent in the array). Conversely, accessing a large character set (like Unicode) for non-CJK scripts (languages using the Arabic, Armenian, Cyrillic, Georgian, Greek, Hebrew, Indic, Roman, Thai/Lao and other alphabets) is repeatedly using a relatively small collection of characters which are located not too far apart (« dense » access). Dense access of a dense array suffers a much smaller performance penalty for fonts that are not CID-keyed than sparse access of a dense array on most computer systems. This is due to the nature of buffered disk and memory operation.

CID-keyed fonts dispense with the notion of a character name, since all glyphs are referred to by the CharacterID number. This has advantages and disadvantages. The advantage is that politically-incorrect and language-script dependent naming issues are avoided. The disadvantages include the lack of a positively identifiable human-readable name for each glyph and the lack of consistency with previous PostScript formats, which used character names like 'ampersand'. Creating new CID-keyed fonts using « borrowed » characters from other CID-keyed fonts is also much more complicated. Finally, calling characters from a font in PostScript by character name only (overriding a font encoding) is impossible, unless an extension to PostScript is made in the future.

(another definition follows)

ClearType technology makes no improvement on cholesteric LCD's, because the color elements are stacked. However, cholesteric LCD's can be made physically much smaller, which allows much higher resolution displays exceeding 200 dpi to be manufactured. [200 dpi is approximately twice the average resolution of computer color display devices of the late 1990's.]

Any number of CMAPs can be used for a given font. However, a CMAP must be composed based on the specific character content of a CID-keyed font. If a number of CID-keyed fonts have the same character content, a given CMAP can be used with all of them.

Production First Software has extended this concept using the LanguageGroup Kit to enable codepages which span more than one font and can address more than 256 characters. The Production First Software implementation of codepages is not at the system level; and more than one codepage may be used on the same document page.

TrueType font formats allow inclusion of a list of codepages which the font may support. At least one application (Microsoft Word97 and beyond) makes use of this data. However, since codepage definition and support is usually implemented at the system level, it seems unwise to evolve a system where the only codepage support from an application is triggered by the codepage list built into a font. With this methodology, if additional codepages are designed and implemented in an operating system, preëxisting fonts would not have the information built-in, even if the character content of those fonts would support the new codepages. A better methodology is to be able to install codepage support through the application, with partial or full support by any particular font depending on its character content. The Production First Software LanguageGroup Kit allows this for PostScript fonts with applications like PageMaker, without the necessity of building in any data about codepages in the font itself.

For 1-byte encodings and codepages, this is indicated here as <nn> (for example <a8>) or 0xNN (for example 0xA8 ).

For 2-byte encodings (such as Unicode or ISO/IEC/10646 in UCS-2 format, U+NNNN or <nnnn> is used (an example being U+20A8 or <20a8>).

For 4-byte encodings (such as ISO/IEC/10646 in UCS-4 format, U-NNNNNNNN or I-NNNNNNNN or <nnnnnnnn> is used (an example being U-005D20A8 or I-005D20A8 or <005d20a8>).

The notation convention used in this Encyclopædia is usually the < ... > forms with lowercase letters in the hexadecimal numbers. Occasionally, the U+, U-, or I- forms with UPPERCASE letters in the hexadecimal numbers will be used.

encoding vector.

COLD   An acronym for Computer Output to Laser Disk. A methodology of storing data.

cold type   Used to designate type set by photocomposition or other computer-driven typesetting methodologies (including desktop publishing). The alternative is « hot type, » which usually refers to generating cast metal type.

CoolType   An Adobe technology announced in 1998 combining a PostScript 3 rasterizer (or parts of it); and Type 1, OpenType, and TrueType font and printer driver interfaces. The areas of use include page layout applications, operating systems, and related font utilities. CoolType could be considered a replacement for Adobe Type Manager (ATM). The first CoolType software was incorporated into a page layout application called InDesign in late 1999.

colonial typeface   A somewhat informal classification of typeface designs which have rough outlines to simulate poor casting techniques and poor printing results common in the U.S. colonial era. Examples include: Caslon Antique, MisionPF Antique, Old Claude, Poliphilus, and 16th Century Roman.

color calibration   See GATX, GetragMacbeth, or IT8.

colophon   An inscription usually placed either on or right after the title page, or at the end of a book, document, paper, report, (or, now, Web page) which provides information on the calligraphic or typographic composing details.

color gamut   The spectrum of colors that a device or system can reproduce.

color palette or color pallet   A collection of specific colors that are available for an image in a digital representation. For example, the color palette of HTML-related image formats consist of as few as 216 colors; whereas the color palette representable in 8 bit color is 256 colors; in 24 bit color is 16 777 216 colors; and in 36 bit color is 68 719 476 736 colors. Generally speaking, if a color is specified which is incapable of being represented by an application, output device, or other rendering system due to being excluded from the available color palette, it usually is rendered by dithering.

Continuous tone images (such as those made by photography) display patchy areas of color rather than continuous tones if millions of colors are not reproducable.

color or color space   The collection of all possible colors reproduceable by a specific environment or recipe.

Specifically, with no adjectives in front of the term, color space implies the complete collection of colors in the spectrum of white light. When adjectives are applied (for example: CIE color space, CMYK color space, Land colopr space, RGB color space), the term denotes all colors reproduceable by that particular color reproduction technology.

COM   An acronym for Component Object Model Architecture, a client/server communication protocall used in middleware, architected by Microsoft. A competing protocall is CORBA, an international standard.

combining character   A character whose character glyph usually is used along with another character's glyph (which character may or may not be a base character) and combined into a single glyph. The distinction between using a combining character glyph along with another character glyph, and a ligature is that in the former case, the combining character glyph usually does not entirely give up its visual form. However, when a base character glyph is used to create a ligature glyph, the resulting ligature glyph could be visually quite different from using the composing base character glyphs individually.

A combining character is usually not used alone in regular text, although it could be under special circumstances. When a combining character glyph is used along with a base character glyph, the resulting character is called a composite character.

composite character   A text character whose corresponding glyph is constructed from more than one base character glyph; can be decomposed into more than one base character; or whose corresponding glyph is internally constructed within a font by using more than one character glyph, at least one of which is usually a base character glyph. A composite character is also known as a « precomposed character. » An example of a possible composite character is Á (Aacute), which can be composed of the characters A and ´ (acute). Composite characters in Production First Software fonts can be composed of an unlimited number of base characters. Some ordinary PostScript fonts might define what should be a composite character as a base character. This increases the amount of data unnecessarily.

Certain scripts use a large number of composite characters (with some of the composite characters being ligatures). These include Arabic, Hangul, and Roman. Currently, there is a big controversy as to whether, or how many, composite characters should be placed in multibyte encodings, such as Unicode and ISO/IEC/10646. The alternative to placing composite characters in encodings is to rely on artificial intelligence algorithms in composing software (word processors, page layout programs) or font resources to properly represent composite characters on-the-fly. This strategy is not simple to implement for any script if text justification or track kerning is employed. Even if neither technique is used, it is not simple to implement (in fact, theoretically impossible) for Roman scripts when pair kerning is utilized. The reason is that when a sequence of characters is spelled out to represent a composite character with another character, it is impossible to precisely determine which two adjacent characters are to be kerned without human intervention for each and every character pair. Oftentimes, positioning of base characters to create a composite character on-the-fly requires discretionary information as to positioning. Usually, this is the purview of the typeface designer. Therefore, the text rendering algorithm must obtain positioning information from a data file, which is different for each typeface. Some typefaces use different designs of diacritical marks depending upon the base character; and some typefaces use a different design of the base character glyph depending on the diacritical mark. These are yet additional permutations which must be recorded as data and then used by the positioning algorithm. For some typefaces, certain composite characters are replaced with ligatures, again at the discretion of the typeface designer. An example of all these effects can be seen in the Production First Software typeface LafayetteU0PF. The use of composite characters eliminates these problems, complications, and obstacles. It should be noted that using pair kerning in document text can increase the comprehension speed by up to 50%, and typically reduce document viewing size (or the number of pages) by up to 30%. Using composite characters could reduce document file sizes by 10% to 30% depending on what percentage of text characters are composite characters.

The disadvantages of using composite characters turn out to be relatively minor. One disadvantage consists of the number of additional code points required in the encodings. This, however, does not translate into larger documents or file sizes unless the composite character additions require more bytes to represent a larger encoding space. Document size is proportional to code point byte width for native representation; but some schemes like UTF-16 can substantially reduce this penalty. The only other significant disadvantage is a possible increase in the amount of kern pair data (if a significant number of composite characters are kerned). Kern pair data is a sparse matrix whose size is proportional to the square of the number of characters kerned; but kern pair data is only font-dependent, not document-dependent. However, most typefaces do not include a significant number of kern pairs having composite characters, except typefaces produced by Production First Software.

combining class   A property of characters, which is a priority number assigned to a character used to specify how a character combines with another character typographically. It is important for using multiple diacritical marks on the same base character, and the use of vowels in complex scripts.

combining mark   A combining character which is a diacritical mark or tonal.

Communicator   A Web browser and Email communication software developed by Netscape Inc. It is available for free from Netscape's Website.

Common Desktop Environment   See CDE.

Compact Font Format   A PostScript, nearly completely binary, font format which allows one or more Chameleon, CID-keyed, MultipleMaster, single, or synthetic fonts to be bundled together in a compact representation. Metrics are handled in the same manner as previous PostScript font formats.

Compatibility Zone   A zone in ISO/IEC/10646 and Unicode which contains character variants in use in other standards which should not or cannot be unified with characters in the standard script blocks.

complex script   A catch-all term referring to scripts which require special processing in typical use (other than a dedicated change in writing direction), relative to plain text, because of some complexity in their nature. Examples include scripts which form conjoins (such as the Indic scripts), scripts which require contextual analysis for glyph selection (such as Arabic, Indic, or Korean); and scripts which may have alternating writing directions or unusual line-break requirements (such as Mongolian).

complex document or compound document   A document composed of text, graphics, images, sound bites, animations, movies, or other non-textual components.

compression scheme   A technique for reducing the size of a file by using a specific algorithm to eliminate some of the data. Depending on the nature of the data and the algorithm, if the eliminated data does not result in any change in representable content (no unreconstructable loss of data representation), the scheme is termed « non-lossy » or « lossless. » If a change in representable content occurs (an unreconstructable loss of data representation), the scheme is termed « lossy. »

composite font   A font whose font definition points to other fonts in a hierarchical manner. While a composite font can be said to be « composed of other fonts, » this terminology is misleading, because composite fonts do not require duplication of data.

conventional font   The term here denotes a single-byte font, or a font accessed by single-byte characters.

core font   A font which is bundled with another software product.

composite font extension   A feature available on some PostScript Level 1 interpreters which allows the use of Type 0 composite fonts. This feature is also known as the « Kanji font extension. »

compound ligature   A ligature which is constructed from a base character and another ligature from an action not within the font.

compressed   Refers to a typeface which is a narrower version of another typeface but with nearly the same vertical stem widths.

Compression schemes are usually encountered with and utilized for audio and image files, and, rarely, for entire text or entire document files. Common image formats which utilize compression include .GIF, .JPEG, .PCX, and .TIFF, although compressed TIFF is rarely used at this time. A compressed file is considered to be « coded, » and a decompressed file to be « decoded, » although this terminology is imprecise. A separate software utility which perform these functions is sometimes called a « codec. »

A number of compression schemes are known, and some are commonly-used: ASCII-85 (PostScript), Discrete Cosine Transform (JPEG, PostScript, TIFF), Fractal, Huffman (CCITT Group 3, CCITT Group 4, JPEG), JPEG compression (JPEG, MPEG-1, MPEG-2), Lempel-Welch-Ziv (GIF, TIFF, .ZIP files), Run-length (Dr. Halo, Macintosh Paint, Windows Paint, PhotoShop, PostScript, Sun Raster Data, SGI, Targa), Stuffit (.SEA and .SIT files), Wavelet, Zip (.ZIP files), and Zlib (.PNG files).

Computer Style   A typeface structural style having the characteristics of simulating computer-generated output, such as printout, ocr, and illuminated displays.

comstock   Same meaning as inline.

condensed   Refers to a typeface which is a uniformly narrower version of another typeface either by design of the font or by manipulation from within an application.

conjoin   A property of glyphs representing a string of characters in certain alphabets joining together before being represented or presented in visual form. Arabic and Indic alphabets require horizontal or vertical conjoining in correct typographic representation. The noun describing the result of a conjoin is a « conjoint » (also known as a « conjunct » ).

Conjoins are often implemented in operating systems by simply placing a conjoined character glyph in the proper position, overlapping it slightly with adjacent glyphs in some cases. There is a problem with this type of implementation: if an application specifies outlined or textured conjoined characters, this cannot be done properly because the conjoined glyph outlines are not continuous. Conjoined glyphs must then be reconstructed with continuous outlines, thereby removing overlapping outlines, usually by a graphics application. An alternative treatment is to replace the conjoined pair with a ligature; but this has practical problems in that the number of ligatures required are equal to the number of pair combinations of characters which conceivably could conjoin. Conjoint ligatures are common in Indic scripts, because the conjoined forms are visually different than the isolated forms. Conjoint ligatures are less common in Arabic, not being used in everyday text communications.

conjoint   See conjoin.

conjunct   See conjoin.

Conseil Europeen pour la Recherche Nucleaire   Also known as « CERN » and « European Laboratory for Particle Physics, » the scientific institute located in France and Switzerland which performs nuclear physics research on fundamental nuclear particles, the building blocks of nature. Much of the work involves designing and operating particle accelerators ( « atom smashers » ) to produce a variety of heavy particles for study. In 1990, CERN developed the World Wide Web, a graphical extension of the Internet.

contextual commerce   A construction strategy of Web sites in which the opportunity for transactional processing is only associated with certain, specific elements of page content. A simplified, hypothetical example would be a link to a transactional page for the purchase of specific products included in a list of suggested products, some of which would be free.

contextual editing   The checking of data for correctness. A spell checker is an example of contextual editing. Another example is form validation in HTML, wherein form data is checked to determine if it fits within limits for the particular item of information collected.

contextual processing   The task of selecting (other than by simple character/glyph mapping) the correct glyph or shaping a glyph on-the-fly to represent a text character and determining the layout location to render it. This process can be optional and relatively simple in some scripts like Greek, Hebrew, or Latin (for ligature substitution or for substituting a medial form), or it can be complicated for scripts like Arabic (selecting one of several forms and adding kashida glyphs), Egyptian hieroglyphics (mirroring and selecting different forms), Indic (selecting different forms, relocating and joining glyphs), and Mongolian (selecting different forms, joining and relocating glyphs, modulating the writing direction). Scripts requiring these capabilities are said to possess contexual sensitivity.

In Arabic, Armenian, Cyrillic, Georgian, and Latin, ligatures must be specified based on a sequence of more than one character. Ligature substitution is a relatively simple contextual processing task. For example, in Latin, the ligature glyph fi replaces individual glyphs for f and i. In Arabic, the proper form of the glyph (initial, medial, final, or isolate) must be determined by analyzing a text string representing a word. In hieroglyphics, it is necessary to determine whether the right or left hand glyph image is required. In both hieroglyphics and Mongolian, the writing direction of a particular text string must be determined. In Indic, the proper glyph must be selected based on the placement; and the placement (above, below, or beside the previous glyph) must be determined based on other characters in the text string. A script that requires contextual processing in typical use is usually called a complex script. Contextual processing is usually performed by an input method editor.

contour   Another name for inline.

contraction   Strictly speaking, in character space, the replacement of more than one character in a string by a single character. This could be considered (only improperly) an analog of a ligature in glyph space because the replacement (character or glyph) or the objects consumed (characters or glyphs) may not correspond.

contrast   The change in stroke thickness between the thickest stem or curved stroke portion and the thinnest stroke or thinnest curved stroke portion.

A typeface is said to exhibit high contrast when the aforementioned difference is greatest. Typeface examples of this are Bodoni Poster and Thorogood Roman. The opposite extreme is a typeface with no contrast. Typeface examples of this are BernalPF and Courier.

Copyright Convention   See the Berne and Pan American Copyright Convention.

CORBA   An acronym for Compliant Object Request Broker Architecture, a client/server communication protocall used in middleware. A competing protocall is COM, (architected by Microsoft).

corner mark   A right angle-shaped mark used as a registration or cut mark.

Corporate Use subarea   A portion of the ISO/IEC10646/Unicode encoding Private Use area starting from [U+F8FF] and extending to lower code points (but not lower that [U+E000]). See also End User subarea.

counter   The open area between vertical stems in a letterform.

counterform   A term sometimes used to denote the space between letterform parts enclosed by an imaginary rubber band stretched around the outside of the letterform.

country code   A code used to specify a particular keyboard character set and layout for the language of a particular country.

CP   An abbreviation for Code Page or codepage.

CPI   An abbreviation for Characters Per Inch. See pitch.

cpp   An abbreviation for characters per pica.

CPSI   An abbreviation for Configurable PostScript Interpreter. A PostScript interpreter which runs on a host or server microcomputer system, generates bitmap data, and sends the bitmap data to an output device which serves strictly as a marking engine.

The advantages of CPSI is that it is easily upgradeable and it has direct access to font resources and input files stored on the system hard disk or other mass storage devices. The main disadvantage over embedded (firmware) PostScript interpreters is, so far, throughput. Throughput depends on high speed processing and the ability to quickly send huge amounts of bitmap data to the output device. The processing speed penalty occurs because specialized firmware can be highly optimized. The processing speed issue, however, will improve as microprocessors get faster, which is happening almost continuously. The penalty of sending huge amounts of bitmap data to the output device can only be reduced, however, if faster data transfer protocalls are designed. The data transfer rates are affected by the system bus and disk subsystem interface. These protocalls cannot improve in speed almost continuously like microprocessors, because they make use of standardized specifications, which do not change frequently.

Cray Research   A company spun off from Control Data Corporation (CDC) of Minneapolis, Minnesota, USA in 1971, made famous by developing commercial scientific supercomputers in the 1970's-80's, including the Cray-I, Cray-I-XMP48, and the Cray-2. The Cray's were characterized by a circular CPU compartment approximately 7 feet wide and 6 feet high, surrounded by a circular arrangement of cushioned seats, and weighing approximately 8 tons, because there were so many semiconductor memory units packed so densely. The resulting short electrical connections added to the speed of the processor. The original Cray-I was a single processor system, but the Cray-I-XMP48 and Cray-2 were multiple processor systems with user-programmable parallel computing. The Cray's originally ran a Control Data Corporation (CDC) operating system, their lineage and architecture somewhat related to the CDC-6600 and 7600 developed by Seymour Cray, who later left CDC to found Cray Research. Varieties of Unix later ran on the Cray's. The Cray's were mainly used for compute-intensive monte carlo and deterministic mathematical modelling tasks in such fields as nuclear weapons design, nuclear fusion research, particle and radiation transport, aerosol transport, weather forcasting, molecular chemistry and pharmocologic research, fluid dynamics, structure stress design calculations, earthquake prediction, and movie animation.

With the advent of ever-faster microcomputer chips, the speed advantage of the Cray's was continuously being eroded, and the expensive operation (air conditioning, electricity, space) helped the demise of that technology. Cray Research, however, continued to develop supercomputers using microprocessor chips in massively parallel processor (MPP) configurations. In 1995, a Cray-I computer, retired from Lawrence Livermore Laboratories in Livermore, California, was advertised in the San Francisco Examiner for sale as scrap. It was purchased by an entrepeneur for US$10,000. In 1996, Silicon Graphics Inc. (SGI) purchased Cray Research for US$740 million, as well as Paragraph International in 1997, a Russian software company owned by world chess expert Gary Kasparov, who later lost a widely-publicised chess tournament to IBM's MPP RS/6000-based supercomputer. The purchases nearly bankrupted SGI, sending SGI stock plummeting from US$25/share to US$8/share, having not yet recovered by 2000. In January 2000, SGI sold Cray Research for US$100 million to Tera Computer Company of Seattle, Washington, USA. Tera was founded in 1987 by Jim Rottsolk and Burton Smith to develop supercomputers.

crawling robot   A software utility, not under the control of an end-user, that examines a number of Web sites to determine if a specific action is to be taken. Crawling robots take a variety of forms:
1) Information gathering and servicing tasks, such as Web shopping;
2) Data mining tasks, such as looking for data patterns;
3) Monitoring of specific Websites, such as a search engine looking for sites to add a homepage URL to a list for later use; or to establish netcasting from a site.

.CRD file   See bitmap formats.

.CRE file   See bitmap formats.

CRM   An abbreviation for Customer Relationship Management. A customer retention strategy of setting up a centralized database that customer support and customer call centers can access from any place in the world.

crossbar   A generally horizontal stroke which completely intersects a generally vertical stroke. Examples can be found in: A, T, f, t.

cross-format   A designation given to a software product package which contains versions of the product in different formats, or one version which acts as if it were equivalent to versions of different formats.

cross head   A heading interposed within body text for separating or identifying different subject areas of the text.

cross-platform   A nebulous term which can imply consistency of architecture in the issues of data exchange, or identical functioning, on different platforms. « Cross-platform » does not imply « platform-independence », where the same identical software product can function on different platforms.

cross-script   A term designating a property which spans more than one script. An example would be cross-script pair kerning, which is present in most Production First Software Unicode and UCS-4 typefaces.

cross-script kerning   This refers to pair kerning where the characters of the pair belong to different scripts. Combinations include, but are not limited to, Cyrillic-Latin, Cyrillic-Greek, Greek-Latin, and Hebrew-Latin. This arises due to hybrid alphabets, which are used in some African, eastern European, and native American languages.

crossware   A term used for electronic communication software which enables cross-communication between Internet and intranet systems.

crotch   The area immediately above the junction of the two diagonal stems of the letter 'M'.

CRT   An abbreviation for Cathode Ray Tube. A device for displaying images consisting of a large glass container, containing a vacuum, with a phosphor screen at one end and a high-voltage electron gun at the other end. It is used in televisions and desktop computer monitors.

CTP   An abbreviation for Computer-To-Plate. See platesetter.

.CUR file   See bitmap formats.

cursive   See script.

(another definition follows)

cursive or cursive writing   A flowing style of writing (« handwriting ») where letterforms are continuously linked.

Some primary schools are no longer teaching cursive writing, because there are questions about its importance in the computer age, where keyboard buttons, mouse clicks, and touch-sensitive screens are used to create text.

cursor   A moving position indicator on a computer monitor screen.

CUS   See Corporate User subarea.

Cyrillic Extended   A Production First Software supplementary Cyrillic and Old Cyrillic character set which includes historic and religious variants, ligatures, numerals, upper and lower case diacritical marks, and symbols not in the Extended Cyrillic character set.  
 
 
 
 
 
 
 
 
 
 
 
 
ERROR

Production First Software  |  INDEX  |  SEARCH  |   PREV   |   NEXT   |   HOME   |   HELP   |  Copyright & Disclaimer Notices