Terminology

 
ASCII/ANSI
a standard (one byte) character code used by MS-DOS (ASCII) or Windows (ANSI)
attribute
something that signifies a tag or element
class, class name
a class subsumes a specific set of instances like the class MONTH subsumes January, February ... December.
command file
a file containing the parameters for driving an application
ECI
acronym for EPTAS Client Interface
element
see tag
end tag
determines the end of a tag; written </any-string>
EPTAS Client Interface
MS Windows application for accessing EPTAS applications
exact match
a LRD match where the source LRD entry matches immediately 100% with the source document match without having to apply replacement classes
full match
a LRD match where the source LRD entry matches 100% with the source document match after having applied replacement classes
HTML 3.0
acronym for Hypertext mark-up language; the tte format uses most of the defined tags of HTML to mark text format information.
LinguisticObject
a basic object containing monolingual information about a term; LOs usually reference MATs; LOs are referenced by loids, which in turn consist of a matid plus a language code plus a number. The matid can be deduced from the loid.
LO, MOL
see LinguisticObject
LRD
acronym for Linguistic Resources Database
MAT, MUL
MultiLingualAlignedText; a basic object holding the translation relationship between different Los like source language and directionality; MATs are identified by matids.
Tte / tte format
standard text and workflow description format for EPTAS; based on HTML 3.0
reference
In some cases objects like LOs are not directly stored in their „father“ object (e.g. a MAT), but the father object only contains an address pointer where the actual object (the LO) can be found (read from the database).
regular expression
An expression which consists of fixed and / or variable parts, like *, ?, +; e.g. ab(.+)cd; they are matched against a string like ab34cd.
replacement class
a special class (e.g. month) which can unique be transferred into its target language counterpart. Replacement class could be translation classes like month (Mai - May) or something like numbers ... Replacement classes are used to improve the quality of a match by translating strings which could be attached to a replacement class. RiskLevel can be associated with replacement classes defining the safety of a replacement class.
SGML
standardised generalised mark-up language; description for text structures based on a meta language
tag
called element too; associates a symbolic name with values which themselves can contain tags as values; is written in <any-string> brackets; usually contains a value and sometimes attributes. Example: <Match KeyDifference=20>My small example</Match>
TM
an application which translates texts by searching for similar sentences in a database
Phrase Translator
an application which translates texts by replacing sub-sentence parts
UNICODE
a standardised two byte character code; UNICODE is a compatible character set to ISO 10646 - UCS. As UNICODE is better known we use this term instead of ISO 10646 which would be the correct name.
variable
a part of the a string which can be substituted by another value; within the Phrase Translator Importer variable parts are designated by [<variablename #no=<no>]. The variable name designates a class (class name) like NUMBER, MONTH ...; with regular expression matching a variable consists of “\” plus a variable number, e.g. “\1”. It matches a variable part of a regular expression, which is designated as “(...)”.

Feedback: Dr. Klemens Waldhör
Project Site Responsible: Dr. Klemens Waldhör Site last updated: 05.04.2000
Proposals for Site Changes: Dr. Klemens Waldhör