Text in Multimedia

• Definition: the medium of delivering information via words to be read and interpreted by the reader.
• Types of writing systems:
• Logographic - is a single grapheme (anatomic unit of written language) which represents a word or a morpheme (a meaningful unit of language).
• Alphabetic - using a set of standardized letters representing phonemes (smallest contrastive unit in the sound system of a language).
• Syllabaries - using symbols to represent syllables.
• Ideographic (or pictographic) - a graphic symbol representing an idea.

History of Text

• Cuneiforms - Sumerian, 4-3k BC.
• Hieroglyphs (using pictures to represent ideas and objects). Various existing types - Egyptian (~3k BC), Anatolian (Luwian, ~2k BC), Cretan (~2k BC), Mayan (3rd - 16th century)
• Alphabets: around 1700 BC, the Semite population in Egypt developed an alphabet that is believed to be the ancestor of most modern ones, although it's not itself understood.

Current Writing Types

• Alphabets: Latin, Cyrillic, Arabic, Brahmic, Hebraic, Greek, other.
• Syllabaries: small scale used in Greece (Mycenaean), Native American (Cherokee, Cree), African (Vai), China (Yi). Large scale: Japanese kana (hiragana and katakana)
• Logographic and ideographic: Chinese, Japan (Kanji), Korea (Hanja).
• Spelling: phonetic alphabet (international). Most languages will have digraphs (2 letters), trigraphs or even tessergraphs (tsch) to represent one phonem.

Alphabet distribution in the world

Cree Syllabary

Typeface and Font

• Typeface - family of graphic characters including many sizes and styles. Example: Arial.
• Font - collection of characters of a single size and style belonging to a typeface. Example: Arial 14-point bold.
• Size - expressed in points (0.0138 in). A 14-point size means from the top of capital letter to the bottom of descending letters (like y).
• Case: uppercase (capitals) and lowercase (small letters). Case-sensitive: any system that distinguishes between the uppercase and lowercase.

Font Terminology

• Bitmap font - each glyph is stored as a matrix of pixels. Harder to scale (need anti-aliasing) or to condense and expand.
• Outline fonts - the glyphs are stored as an outline using vector graphics (PS, TT).
• Postscript fonts - Type 1 or Type 2 fonts, developed for the PostScript (printing oriented) language.
• TrueType - developed by Apple & Microsoft, higher degree of control of how the fonts are displayed.
• Serif - small decorative line added to the font.
  sans serif (Arial).
• Monospace font - a font for which all the characters have the same size. Useful in situations where the alignment is important (programming code, ASCII art).

Character Encoding

• ASCII - American Standard Code for Information Interchange. It represents a character as 1 byte (8 bits). This means up to 256 characters. Traditionally 128, extended 256.
• Unicode - Universal Character Set represents each character on 2 bytes (16-bit), extending the range to 65,536 characters which can include many alphabets and notations.

Hypertext

• Hypermedia - interactive multimedia enhanced with links between elements that allow the user to browse the project.
• Hypertext - a hypermedia project where the text component represents an important part.
• For collection of documents containing text beside the user-defined links, textual information can be processed and indexed.
• This makes the retrieval of individual parts of the collection easier using a search engine based on keywords.

Indexing

• The documents to be indexed are first separated into words.
• The most common words are removed (as, of, are).
• The remaining words are processes to remove suffixes and prefixes like s, ing, ly, etc. This turns several words from the same family into the same thing (work, works, working -> work), called stems.
• The stems are counted in each document and some importance is decided for each of them.
• An inverted index is generated, storing for each keyword (stem) the documents (pages) that are related to it.

Search Engine

• The user enters one or more keywords to search for, with or without Boolean operators (and, or).
• These keywords are processed the same way as described for the indexing.
• All of the documents/pages related to the processed keywords in the inverted index are retrieved and classified by relevance.
• Search engines use various techniques to sort among the retrieved pages. Hits from the user on a page can improve its relevance or links from a page to another can also make it more important.